A New Approach to Understanding the Nature of Word Meaning


.



Summary.

The nature and constitution of word meaning is one of the prime puzzles of linguistics. What
is the cognitive content of word meaning? What do we know when we recognize a word, that
enables us to understand it in an instant? What enables a word to refer toentities and events
in the world? What enables words to operate in the way they do in forming sentences? Most
linguists assume that word meaning consists, at least in part, of constituents known as
semantic features, but no means of identifying them in a definitive way has been found. This
paper describe such a means and the features that have been found. It presents a small set
of semantic features that can be identified with a high degree of confidence by a surprisingly
simple process that has been overlooked. These features combine to form the structural
framework of the meanings of words. In terms of the title of this paper they constitute an
‘alphabet of word meaning'. They provide the necessary structure, but not the full content of
the meanings of words in any language. The set of features has been named semantic
factors
. They are also present non-linguistically in how we know and experience the world
and how we respond to it. There is reason to believe that the factors are profoundly biological.

A key aspect of the semantic factors is that they appear to correspond to a range of
body-related (somatotopic) regions of the brain thus creating a neural code that translates
directly into specific, structural semantic features, the semantic factors. In this way the factors
can be seen to form a bridge between neural activity and word meaning. If this conclusion is
sound the factors are an unrecognized property of language with significant explanatory
potential for some central processes of language that have been difficult to explain.

The paper concludes with a comparison of its approach to language with that inherent
in Chomskyan Generative Grammar. Chomsky proposes a biological basis of syntax that is
manifested in a Universal Grammar. The paper's approach, on the other hand, involves a
Universal Semantics of the word that is arguably central to the wider operations of language
and may have implications for the Universal Grammar.

The paper is a work-in-progress, but if my conclusions about the nature and function
of the semantic factors are validated by other authorities they may begin to provide a new
dimension for our understanding of language. The purpose of the website is to make the
results of my research available to anyone who is interested in understanding the operation of
word meaning, but especially to other language researchers. Unfortunately, ongoing attempts
to interest linguists in the results of this work have been unsuccessful for a number of reasons
such as the factors' lack of fit with some central precepts of linguistics, the fact that I am not
an academic, and the source of the discovery – an insignificant backwater of linguistics.
These have precluded any significant interest from linguists, let alone peer review or
collaborative research, so far. My position as a non-academic claiming to have identified a
central aspect of language that has largely escaped the attention of linguists is unusual, to
say the least. My claims about the implications of the semantic factors need to be judged by
the evidence and arguments that are presented in the paper.

An Annotated Contents.

As an introduction to a newly developed aspect of language this essay is necessarily
quite long and detailed but it is not intended to be read from beginning to end. The
large Sections 6 and 7 in particular, are a resource to be sampled. The lengthy
Section 15 may not interest some readers. I am providing a detailed contents list to
enable readers to navigate the document by identifying sections of particular interest
to them. click on the headings below to go to each section

1. The nature of word meaning.

Outlines the paper's central issue, the determination of the nature, constitution, content and
source of word meaning. These are are poorly understood currently. Semantic features,
facets of the meanings of words, are assumed by most authorities to be key constituents of
word meaning. The newly discovered semantic factors are a type of semantic feature that
has some very surprising and unique characteristics and a powerful explanatory capacity.

2. The inscrutability of word meaning.

Describes the difficulty in investigating word meaning and its comparative neglect in recent
linguistics. The embodiment of word meaning, is identified as central to the approach
developed in the paper.

3. An unrecognized property of language.

Describes a proposed previously unrecognized property of the human language faculty – its
ability to access primal, abstract elements (dimensions) of cognition that arguably form a
bridge between the neurons of the brain and constituents of word meaning. This property is
manifested in the semantic factors. The factors operate in language as an ‘alphabet of word
meaning' that establishes the necessary structure of words' semantic content at an abstract
level. The identification of the factors is outlined in Section 5. An initial example is provided of
how the semantic factors operate in the meaning of a word. If this hypothesis is validated it
offers an entirely new way to understand word meaning.

4. The curious world of the phonaesthemes.

Phonaesthesia, a form of sound symbolism, is responsible for an important segment of the
English lexicon (such as many of the cl-, gl- and str- words). The section provides a
description of four phonaesthemes. The full 34 word-initial phonaesthemes of English are set
out in Annex 2. These sub-morphemic constituents of words afford a preview of the following
discussion of the semantic factors. There are striking similarities and differences between the
two. Phonaesthesia is analysed in some detail to provide a backdrop to the presentation of
the semantic factors in the next three sections.

5. A new form of sound symbolism: articulatory iconicity.

This section outlines the broad characteristics of the semantic factors that distinguish them
from the phonaesthemes and other proposed types of semantic features. It describes how
these much more elusive entities were identified and explains why they were not identified
earlier. The articulatory iconicity (similarity of forms of articulation of sounds and the factor
meanings) that enabled the factors to be identified is discussed.

6. The meanings of vocal sounds: the descriptive semantic factors.

Lists of core words (the most frequently used and closest to basic-level human experience)
from the relevant sections of English dictionary are used to obtain the twelve descriptive
factors. The semantic factors are of two main distinct types. This group represents some key
physical parameters of the world (e.g. surface, particularity) that are also constituents of word
meaning (as in face, floor, point, put). The iconic motivation of each factor, based on the
initial phonemes of words, is described. This iconic association of sound and meaning is
prominent in every section but one. Although this iconicity enabled the factors to be
discovered, they occur in words as constituents of meaning far more frequently without any
iconic association with sounds.

7. The meanings of vocal sounds: the affective semantic factors.

The ten affective factors are described in the same way. These represent a totally different
but complementary aspect of word meaning, affective dispositions that are common to
human experience and to the meanings of words (e.g. positiveness and negativeness).
These factors have a central function in the meanings of many words. Sections 6 and 7
contain the empirical evidence that provides the basis for the rest of the paper. Having
identified the factors in this way, they can be seen collectively to constitute a newlydiscovered,
centrally important element of language, a Rosetta Stone perhaps, that enables
us to interpret the semantic substance of word meaning in a reliable way for the first time.

8. The semantic factors in Maori words.

An investigation of Maori has revealed the extensive presence of factoral iconicity in most
sections of the Maori dictionary with much the same sound/factor associations as in English
despite the small inventory of sounds in Maori. This is a very surprising result that supports
the validity of the factors and the existence of articulatory iconicity in at least some other
languages. The occurrence of the same semantic factors in Maori and English words with
equivalent meanings is necessary, but the associations of the same and equivalent sounds
with the factors is surprising.

9. A new model of word meaning.

Describes a model of word meaning with the semantic factors at its centre. This is a 3-part
model, the other two parts being a semantic gestalt, the immediate intuited ‘feeling' of the
meaning of each word (an image in some instances), and encyclopedic meaning, the
various types of stored knowledge associated with a word that contribute to its understanding
and its fit in different contexts. The concept of semantic isomorphism is introduced. Gestalts
have an isomorphic relationship with what they represent. This aspect of gestalts is
derived in part from the isomorphic relationship their constituent semantic factors have with
the ontological features of the world that they represent. This isomorphism is derived from
the somatotopic neural basis of gestalts and the factors (see next section).

10. The nature and origin of the semantic factors.

The semantic factors have some unique characteristics that enable them to perform their
function of providing the necessary structure of words in any language. Section 10.2
introduced the new property of language that the semantic factors represent. The factors are
instrumental in enabling language to denote entities in the world. Their ability to do this stems
from their intrinsic associations with five broadly defined somatotopic (body-related) regions
of the brain that enable the factors to operate as a bridge between the neurons of the brain
and language. The factors translate the language of the neurons into a form accessible to
human languages.

The evident primal biological nature and origin of the factors supports this description.
Their function in language is evidently derived from their prior function in regulating the
behavior of human beings, animals and organisms. The alphabet of word meaning is the
alphabet of the behavioral aspect of life in all its forms. This is an unexpected and
extraordinary conclusion. The dimensional nature of the factors is described.

11. How to spell the meanings of words.

The semantic factors, as the alphabet of word meaning, form configurations that constitute
the necessary structure of words' meanings, especially the core words. Two other semantic
elements are required to express the full meanings of words. These are described. Word
meanings are set out in a formulaic manner that identifies constituent factors and the
other elements. Numerous examples are given. The substantial degree to which they provide
effective structures for most core words is compelling testimony to the factor' validity as the
most primal constituents of language. The constitution of word meaning is not uniform across
the lexicon. The section identifies four types of word meaning in which different
proportionalities of the three elements occur. An impressive result is the model's success in
describing the semantic content of more and less abstract words and closed class
grammatical words that are beyond the reach of other approaches.

12. Theories of word meaning.

A number of the main theories of word meaning from the last half-century are described and
evaluated using knowledge of the semantic factors and the 3-part model. These theories are
diverse. Semantic features and encyclopedic meaning are widely acknowledged as important
elements of word meaning, but there is little agreement on their nature and identity. Arguably,
none of the theories have empirical credentials that match those of the semantic factors and
the 3-part model. An outstanding advantage of this approach is its use of the empirical
evidence available from articulatory iconicity for the existence of the factors that are its core.

13. A new theory of word meaning.

The theory inherent in the 3-part model of word meaning is decribed and evaluated
according to a set of desiderata that corroborate its potential to contribute to linguistic theory.
The theory appears to provide an account of word meaning that is more effective than any
other existing theory.

14. Neurophysiological research into word meaning.

Neurosphysiologists have been very active in investigating the neural basis of word meaning
but progress so far has been limited. This research is seriously constrained by the failure of
linguistics to provide a valid theory of word meaning that could guide experiments and the
interpretation of data. There would appear to be major potential for the semantic factors to
assist in this work particularly on account of their somatotopic nature, their systematic
correlation with specific regions of the brain.

15. Implications of the factors for linguistic theory.

This, the longest section, deals with a range of issues. First, it introduces two new metaphors
to illustrate the apparently central function of the factors in language: they can be viewed as
the source code for the basic neural computational system of language at the lexical level;
and they can be regarded as the DNA of language, the instructions for writing the semantic
information that language conveys. If the paper's description of the semantic factors and their
function in word meaning is valid the implications for linguistics appear to be substantial. In
the first instance the innateness of the factors resolves the debate about whether word
meanings themselves are innate as some theorists propose. Meanings are not innate but
their essential constituents are. The section explores possible implications of the factors for:
the emergence of language; the acquisition of language by young children; the relationship of
lexical semantics and syntax with reference to Chomsky's Minimalist Program and the key
linguistic process of sentence formation. Several aspects of the Program are criticized. Key
elements of language that have been attributed to syntax are more validly viewed as
attributes of word meaning. The dominance of syntax and the centrality of logical form in
Chomsky's approach are seen to be challenged by the ‘bio-logic' of the factors. With regard
to the composition of sentences, a lexicalist (word based) alternative to Chomsky's key
element of Universal Grammar, Merge, is advanced. These conclusions challenge an
enormous body of work in generative grammar. The section concludes with discussion of
another linguist with some views on Minimilism similar to some outlined in this paper.

16. Conclusion.

In an attempt to clinch an argument that lacks direct physical evidence – the proposal that
the semantic factors are valid as fundamental constituents of language – twelve matters that
serve to corroborate the actuality of the semantic factors are spelt out.

References.

Annex 1. A full description of the set of semantic factors.

Annex 2. The full set of phonaesthemes in English.




annotated contents




1. The nature of word meaning.

The subject of this paper is word and concept meaning and particularly the nature of
their meaning in our speech, our minds, and our brains. We all know most of the time
what words mean and when we don't we consult a dictionary. From that point of view
word meaning seems straightforward. But there are deeper issues that relate to what
constitutes the knowledge that occurs in a microsecond when we understand a word.
What is word meaning in this personal sense? What is the mental content of a word
as we use it? How is it formed in the neural matter of the brain? What enables words
to contribute to the meaning of sentences in which they occur? What enables these
mental entities to refer to things and events in the world? These are large issues for
which linguistic theory currently has no satisfactory answer.

In comparison with some other areas of linguistics the investigation of the
nature of word meaning has been neglected in modern linguistics. This is for the very
good reason that there seem to be few reliable clues as to what it actually is, in
conceptual and psychological terms, if indeed it is anything more than a learned form
of denoting. We understand words perfectly well in their everyday use. Linguists,
philosophers, psychologists and more recently neuroscientists have been seeking
the answers but there is little consensus except that dictionary-like descriptions are
not the answer. Meanings in the head as we read a sentence or hear someone
speak are something quite different. I propose that some very good clues are
available, but they have curiously been overlooked.

The contents of word meanings cannot be observed directly but various
means of analyzing them have been proposed. A prominent approach has been the
postulation of sets of semantic features, attributes of what a word denotes, that
contribute to its full meaning. This implies an ability to break down word meaning
intuitively into constituent parts. Features may be quite abstract or quite specific or a
mixture of such types. The problem with this approach is that no one has found a
principled means of identifying a workable set of features. Nonetheless, it has been
prominent, in various forms, since word meaning began to be investigated seriously
in the 1960s. For example, Chomsky (1965 and 1995) has assumed that semantic
and syntactic features of words are essential lexical constituents that underpin his
generative grammar in its various forms (Chomsky/McGilvray (2010)) but, apart from
syntactic features, he has concluded that it is virtually impossible to determine what
they are. Jackendoff (2002) and Pustejovsky (1995), two major investigators of word
meaning, have developed substantial and overlapping accounts of word meaning
based on semantic features, but both are speculative. An allied means of analysing
word meaning using semantic features of an entirely different kind has been
proposed by McRea and associates (2005) who postulate a type of feature that
denotes manifest aspects of entities, some common to many words but most specific
to a word, as key constituents of meanings. These are readily accessible, nonabstract,
open-ended in number and are best suited to concrete words. These
approaches are all problematic but the latter has been employed by some
neuroscientists investigating word meaning – Moss et al (2007), Vigliocco (2007) –
perhaps in the absence of anything else being available.

In a number of unpublished papers and in this website paper I describe a
newly identified set of features that can be extracted from the lexicon in a highly
principled way. This is a small set of features that I call semantic factors to
distinguish them from other types of features that have been proposed and to convey
a sense of their active function that contrasts with the passive nature of other types. I
believe the semantic factors shed far more light on word meaning than any other
current proposal. I will seek to demonstrate that the small set of semantic factors are
drawn directly from human experience of the world, that they can be identified
empirically, that they have an evident direct relationship with specific regions of the
brain, and that they have some very impressive characteristics that are not shared by
other semantic features. They are, inter alia, necessary, universal and innate and
they possess a systematic capacity to combine to construct the meanings of words of
all types at an abstract structural level. No other proposed semantic features have
these capabilities.


annotated contents




2. The inscrutability of word meaning.

If anyone could take a stratospheric view of the activity over the last century in the
land of Academia Linguistica they would find an interesting scene, a scene of everincreasing
activity and inexorable expansion into new territories as linguistics
endeavoured to become a true science in full comand of its subject. Most of the
previously well-established areas were being renewed, many totally new areas on the
periphery were being developed and the old province of Grammatica, one of the
busiest, was being constantly and busily reconstructed. However, curiously, at the
heart of the land, the province of Semantica and particularly the district of Verba
Semantica – word meaning – there was much less activity. The workers in this area
that is seemingly important to the whole domain, were finding it difficult to make
significant progress.

Semantics, and particularly lexical semantics, faces an enormous challenge in
meeting the standards of empirical science that contemporary linguistics aspires to
because word meaning is available only though intuition and it is difficult to determine
the degree of consistency of intuitions from one person to another. Intuitions of word
meaning are partly unconscious – we know, but we don't know how we know. But
there are subtle and crucial aspects of word meaning that are ‘sensed' and that can
be brought to mind through an awareness of their nature. One of my main tasks is to
explain how this can be done in an empirically sound manner. (Section 5.3)

A key aim of lexical semantics must be the description of word meaning in an
objective way, but this is exceedingly difficult because such meaning in the
immediate, live sense is, as I have noted, intuited and is variable from person to
person. How does the meaning of full or empty, operate in the mind? – What enables
us to know what these words mean? We have great difficulty in understanding their
semantic contents as they subsist in our awareness no matter how well we know
their meanings. They can only be observed introspectively with difficulty. Word
meaning has seemed impervious to satisfactory empirical investigation. The task of
understanding it appears to be caught in a problematic hermeneutic circle. Hermes
must be proud of this work!

Word meanings vary from person to person, context to context, but the fact
that they are effective communicational tokens guarantees that they have substantial
commonality of meaning, a semantic kernel, as it were. But how they achieve this
has never been satisfactorily explained. They do not seem to be objects that are
suited for scientific explanation. No way has yet been found to extract and observe
this semantic kernel of words that seems to be a curious blend of fuzziness and
ineffable definitiveness.

As a means of penetrating the inscrutability of word meaning I will be seeking
to draw out a feature that is quite prominent in some words but hard and seemingly
impossible to detect in others. As I am using the term, a word's meaning as we
understand it is both mental and embodied. It operates in the ‘mind' but in many
instances it also operates obscurely at a bodily level that we are generally unaware
of. When we bring to mind words like hurt, laugh, wrong and hold, words from the
core of the lexicon, specific complex neural events occur that have some bodily
associations of specific kinds, respectively: sensory, perceptual, interoceptive (gut
feelings) and kinaesthetic. In other core words, such as full and empty, such sense is
more difficult to detect but we will see that this embodied, somatic aspect of meaning
has a crucial role in the understanding how the semantics of the word operates. It is
not something purely mental; it is embodied in ways that prove to be very significant
for gaining a better understanding of language. Embodiment in this context has
become widely recognized but we will see that the semantic factors provide a new
perspective on it.

I propose that there is a way in which we can probe this kernel of word
meaning in the individual mind using a peculiar phenomenon that is common in
English and some other languages. There is a chink in the hermeneutical armour
through which we can gain access to objective information about the actual nature of
word meaning. It is found in an unlikely corner of linguistics, sound symbolism. I will
come to this in Sections 5 to 7. There is also another way that neurolinguists are
exploring––the use of brain scanning technologies. I will discuss later how these very
different approaches may be brought together.


annotated contents




3. An unrecognized property of language.

The proposition that I will seek to substantiate in this paper can be described in the
following way. There is a previously unidentified property of language that allows the
formation of the semantic aspect of words, their ability to denote, in the mind, entities
and events in the world – as well as imaginary entities and events. This applies
equally to concepts. Without this property, it seems, there would be no concepts, no
words, no language. It is crucial to the whole language enterprise. This property of
the language faculty, which manifests itself in the set of semantic factors, is not
originally linguistic. It also operates as a set of basic sensorimotor and motivational
principles of cognition and behaviour – our awareness of key physical parameters of
the world in association with a set of key innate human dispositions that motivate our
interactions with the world around us. Precisely the same parameters and
dispositions underlie human behaviour and govern the meanings of words in all
languages, I propose. The language faculty has appropriated these parameters and
dispositions for the construction of the meanings of its symbols. The new property of
language is its capacity to utilize the semantic factors that are anchored directly in the
body and brain in a way that I will describe later (9.2) to build the necessary structure
of the meanings of words.

The property that the semantic factors provide for the language faculty is their
ability to operate as a bridge between the brain, language and the world in a way that
allows concepts and words to be formed and propositions to be framed. The factors
are dimensions of human somatic activity in the world (as represented in the brain)
and these same dimensions provide the necessary structure of lexical and concept
meaning. They are the dark matter, the Universal Semantics of word meaning.

To give an indication of what the factors are like and how they operate as
configurations in word meaning I will provide a partial factoral analysis of ‘place' as in
‘this is a windy place'. This anticipates descriptions of the semantic factors later in the
paper. There are five kernel semantic factors that operate as necessary structural
constituents of the meaning of this word: materiality, spatiality, particularity,
manifestness, definitiveness, somatic affect
. A place is a material, spatial, particulate,
generally manifest and definitive entity. The presence of the final factor is debatable. I
believe there is aso a subtle affect, felt as a ‘gut feeling', and of variable nature, that
is a necessary part of the meaning of this word. It may derive from the fact that we
often have a proprietory interest in a place. (This mild affect is absent from ‘area').

These factors obviously have a high degree of abstractness and generality.
This is essential for their function of forming the necessary structure of the meanings
of words. This is the rarified atmosphere of lexical structure that the factors create.
The descriptions of meaning produced by this analysis are often more cryptic. In this
case each of them can be seen to be essential for the viability of the word's meaning.
The factors are ‘letters' of an alphabet of word meaning that spells them out to a
variable degree. These same factors, with the remainder that will not concern us
here, can combine in similar kinds of configurations to form the structure of the
meanings of vast numbers of words in all languages in ways that are demonstrated in
Section 11. It will become evident that the factors have a very interesting prehistory in
that they are not exclusively human. They also operate (by and large) in other
animals in motivating their behaviour. They are essentially biological. I will describe
them as the necessary dimensions of the space of the interaction of organisms and
their environments – all organisms, however complex or simple. As dimensions, they
do not require any specific physiological or neural organization or form.

But something more is needed to provide ‘place' with a full unambiguous
sense. There are two further components of word meaning that will be described
more fully in Section 9. The first is what I will call a semantic gestalt, a felt neural
event that is immediately available when we see or hear the word and know what it
means. The gestalt associated with this word with its extreme abstractness and
generality is unusually difficult to specify. In fact it seems to be little more than the
composite of the kernel factors identified above. In many words as we will see, the
gestalt has identifiable additional features, sometimes in the form of a mental sensory
image as in ‘horse' and ‘sharp'. The factors are, as it were, structural ‘dimensions' of
the space in which the gestalt is formed. The remaining element of the meaning of
‘place' is encyclopedic meaning, a conglomerate of information relating to what the
word denotes and how it is used (see Section 9).

These three elements of meaning combine to make ‘place' an intelligible,
coherent and versatile word.

The dimensional semantic factors are of two main kinds. The first are the
human capacities to recognize parameters of the physical world that are critical to
human experience. These include the materiality of objects and substances while
others relate to the topology of objects – the particulateness or specificity that defines
objecthood; the extensional nature of surfaces of things and substances; the rounded
surfaces
that are characteristic of many living forms and of some artifactual objects;
largeness and smallness; the extendedness and contractedness that distinguish
some topologies, and a few others. This small set of ‘dimensions' is arguably
sufficient to form much of the structural frameworks of human cognition of concrete
objects and events.

The other kind of ‘dimension' does not relate to parameters of the physical
world. These are human dispositions, affectivities, gut feelings that are motivators of
action and behaviour and that are essential to cognition. They are affective in nature
and have a distinctly polar character. They include positiveness/negativeness;
fullness/emptiness of value; abundance, availability/scarcity, deprivation; possession,
acquisition/loss, threat; uncertainty/definitiveness; displacement/satisfaction
.

The factors of both types have a distinctive somatic character that is pivotal to
their function. They can be described as sensory (mainly perceptual and haptic
(touch)), motoric, kinesthetic (relating to dispositions of muscles and joints),
proprioceptive (relating to dispositions of the body as a whole) and interoceptive
(affective, gut feelings) and, as such, they impart distinctive somatic characteristics to
many words, particularly the most frequently used words that constitute the core of
the lexicon of any language. These characteristics are readily apparent in many
words once the factors are known but their significance has not previously been fully
appreciated. Less commonly used words with more specialized meaning fall in
another category as described in 9.6.

This aspect of language enables human beings to form concepts and words
that can be manipulated and combined to form sentences, a key prerequisite of
language. However, the fact that these physical parameters and affective
dispositions are shared, in one form or another, with other creatures prompts the
question as to why other animals and particularly our nearest primate relatives have
not acquired language. There are numerous well-documented reasons but one of the
most significant is the inability of other animals to form and manipulate concepts
(denotational neural structures) to a significant degree in the absence of the
immediate stimulus for those concepts. Apes and other animals for example can do
some clever and surprising things but they cannot entertain and combine concepts to
plan some action when the object of the thought is not present. This was a giant step
in the evolution of cognition and language.

As I have noted, there is no general agreement in linguistics about the nature of the
meanings of words, their semantic structure, or even if they possess an internal
structure. Some consensus has formed around the idea that word meanings are
componential or compositional, that they are constituted, at least in part from
semantic features but there is limited agreement on the nature of these components.
I adopt the compositional approach on the basis of evidence that I will present that
we can identify specific semantic components in the meanings of words.

The presence of semantic features has been proposed by a number of
authorities since the 1960s (among others Katz and Fodor (1963), Bierswich (1971),
Weinrich (1971) Miller and Laird-Johnson (1976), Laird-Johnson (1983), Jackendoff
(1983), (2002)), Chomsky (1965, 1995) but their actual nature and the form of their
presence in language has remained highly uncertain. It will be a surprise that a
definitive set of semantic features has been identified in the human lexicon and that
the process of identification is, in retrospect quite simple and eminently replicable.

The significance of the semantic factors is that they establish the structural
framework for the representation, in mental terms, of whatever words denote. This is
arguably one of the most fundamental properties of language. The issue has been
equally vexing in neurological terms: how does the physical matter of the brain come
to denote items and events in the world? Once the semantic factors are recognized
a solution to this problem becomes feasible, I believe.

It will also be surprising to linguists that the new semantic constituents I will
describe were discovered through a previously unrecognized form of sound
symbolism
, the imitation of elements of word meaning by their sounds. This is a quiet
corner of linguists that is of negligible interest to most linguists. It has not seemed to
have any significant connection with the larger issues of linguistics but I will
demonstrate that it holds a route to new insights into how the semantic aspect of
words works.

The factors are not, however, intrinsically sound symbolic. They operate most
of the time without any association with sound. It is serendipitous that their
manifestation in sound symbolism has enabled them to be identified. Before I outline
how the factors were identified, what they are and how they operate, in Sections 6
and 7 and those that follow, I will describe another form of sound symbolism that
provides a useful introduction to the semantic factors because it is much more
familiar. This is phonaesthesia, the association in words of initial and terminal
clusters of phonemes, with specific facets of word meaning. Although the term may
not be familiar, the words it contributues to the lexicon are fominent in all our
vocabularies. Phonaesthesia is a particularly prominent feature of the English lexicon
and is attested in some other languages to various degrees. Although it has been
recognized for many years few significant implications for other areas of linguistics
have been found. I will propose, to the contrary, that it contains clear evidence of the
existence of the hitherto unknown semantic factors.


annotated contents




4. The curious world of the phonaesthemes.

It is hardly surprising that language utilizes a variety of means to enhance its
expressivity in words. Those I have in mind employ some kind of similarity or
iconicity between the physical means of expression––gestures, word forms, the
shape of sentences – and their meanings. There are numerous examples of similarity
at work in language ranging from the sequencing of words in a sentence in a way that
reflects the temporal structure of events to onomatapoeia (Nanny and Fischer
(1999)). One of the most important in English is phonaesthesia which is a major
contributor of words of a specific type to the lexicon. This kind of sound symbolism is
distinguished from onomatapoiea in which the sounds of words imitate sounds in the
world. Phonaesthetic words have a far wider scope of denotation and a complex form
of iconicity.

Phonaesthesia is distinguished by the use, in a group of words, of clusters of
consonants at the beginning of words, and clusters generally consisting of a vowel
and one to three consonants at the end, to express a common salient facet of the
meaning of the words. An example that is often cited is the /gl/ cluster as found in
numerous words that refer to the emission of light in various forms e.g. glare, gleam,
glint, glimmer, glisten, glitter, gloss, glow
. This meaning has been extended in
several directions to words that denote, not kinds of illumination, but a range of
associations with light – illumination, vision and display such as is evident in glad,
glamour, glance, glass, glee, glimpse, glitzy, glory
or even its obverse as in gloaming,
gloat, gloom, glum
and possibly, somewhat obliquely in glide and globe. This
extensivity of the phonaesthemic meaning is significant because the same capacity is
also a feature of the semantic factors, as we will see.

Phonaesthesia occurs, in one form or another, in many languages. It is very
significant in English where it has generated a significant layer of words in the
lexicon. These are words of a specific type – graphically descriptive of sensory and
affective aspects of the world. Most phonaesthetic words do not belong to the core of
the most frequently used words in the lexicon but they form a valuable adjunct to the
core because thay denote shades of meaning that might otherwise be absent. The
full 34 phonaesthemes of English are set out in Annex 2.

I will survey four examples in more detail to draw out the relevant aspects of
this feature of language. The listings below contain five entries for each
phonaestheme: 1. The phoneme cluster. 2. The putative phonaesthemic meaning,
the theme. 3. Some of the most prominent phonaesthetic words. 4. Further possibly
phonaesthetic words allowing for natural extensions of the thematic meaning. 5.
Some common non-phonaesthetic words that begin with the same cluster.
Phonaesthemes often need to be extracted from a much larger group of words that
begin with the same consonant cluster but do not all share the common facet of
meaning. They are nonetheless very distinctive and not difficult to identify in most
instances because of the semantic commonality of the phonaesthemic words.

The meanings given under 2 are my descriptions. A variety of descriptors has
been given by different authorities. The words listed under 3 to 5 are my selections of
the more commonly used words. Phonaesthemes do not always stick to their script.
Their extension beyond their core meaning often makes it difficult to devise a
comprehensive description. Some words may appear inappropriate at first sight
because they do not conform fully with the theme. In any event inclusions and
exclusions at the margins do not affect the conclusions. Core words are in bold. The
first phonaestheme in the dictionary is a good starting point because it manifests all
these characteristics.

1. /bl/
2. defective appearance
3. black, blanch, bland, blank, blaze, bleach, bleak, blemish, blight, blind, blink,
---blob, blot, blue, blur, blush.
4. blade, blame, bland, blare, blaspheme, blast, blatant, blather, blend, bless,
---bleat, blip, bliss, blister, bloat, blond, blood, bloom (v. and n.), blossom,
---blunder, blurb, blunt, blurt, bluster.
5. bladder, block (n. and v.), blow (n. and v.).

This phonaestheme is remarkable in the degree to which it dominates the bl- words if
the words in 4 are included, as I think they should be. It has a well-defined core of
prominent words and quite an extensive penumbra of allied words (4) that are not
such a good fit with the theme although they seem to share something of the
meaning in ways that are sometimes hard to pinpoint. These words diverge from the
primary theme in accord with five principles that are common to most
phonaesthemes: (i) the representation of only part of the theme (blade, bless, blond,
blue, bloom, blossom
represent conspicuous appearance that is not defective); (ii)
transposing the theme into another (non-visual) modality – bland, blare, blast,
blatant, blather, bleat
and others; (iv) a metaphorical extension of the theme, for
example into abstract words – blame, blasphemy; (v) a polar extension of part of the
theme – bless, bliss, bloom, blossom. These are anything but defective.

An important question that arises is, why is this consonant cluster, bl-,
associated with this particular theme, defective appearance? One possibility is that
some feature of how the consonant cluster is articulated is suggestive of the theme
(iconicity), but there is nothing to indicate that this applies here. Another possibility is
that somehow the phonemes of the cluster are responsible for the association. There
is nothing in the current linguistic literature that supports this notion but is decidedly
consistent with an understanding of the semantic factors.

A facet of meaning, a sub-theme, that is specific to all of these words can be
termed, I suggest, display or conspicuous visibility. We will see later that this is one
of the semantic factors that is commonly associated with /l/ by way of an iconic
articulatory connection. The other sub-theme, defectiveness, is also very similar to
one of the semantic factors, displacement, dislocation, being out of kilter that is
frequently associated with /u/. This sub-theme is present in most of these words
although it has no evident association with the factor associated with /b/, bodily
roundedness, convexity
that has a negligible connection with the phonaestheme.

This somewhat confused picture suggests that phonaethemes may be
composites of the semantic factors associated with at least one phoneme in the
consonant cluster, and that the factors may accordingly be basic constituents of
phonaesthemes, at least in some instances. In fact, such a role for the semantic
factors in the English phonaesthemes is supported by a substantial proportion of
them. Reverting to the gl- words of a few paragraphs back, the l factor is very
prominent and the g factor with its meaning abundance, availability is also arguably
operative.

Here is the second example.

1. /cl/
2. associated with a tight grip.
3. clam, clamber, clamp, clan, clasp, class, claw, clay, cleave (stick, cling),
---clench, climb, clinch, cling, clip, clog, close (v. and adj.), clot, cluster, clutch.
4. clang, clap, clash, clatter, clause, click, cliff, clink, cloak, clothe, cluck.
5. claim, clean, clear, cleave (split), clever, clock, cloth, cloud, clover, clown,
---club (a weapon).

The highly physical (kinesthetic) meaning is easily identifiable in most of the first
group of words. Class is in line with principle (iv) above, a metaphorical extension of
the theme – a class is a category that is tightly defined. In the second group (4) most
of the words relate to sounds with a sharp character. This is consistent with the
second principle (transposition into another sensory modality). It seems to relate to
the way the sounds are produced by a forceful contact of two surfaces. This effect
derives from the kinesthetic character of gripping but is quite a different from that of
the first group. In cliff the theme is associated etymologically with climb. In cloak and
clothe the tight grip is loosened to an enclosing. This phonaestheme is a good
illustration of their usual semantic richness.

But is the theme in this case motivated by semantic factors (whether
associated with the two sounds or not)? A case to the contrary can be made, that the
motivation seems to arise iconically from the manner in which the phonemes
combine in their articulation. They are formed in contiguous parts of the mouth by the
action of the tongue on the palate in /k/ and the alveolar ridge behind the top front
teeth in /l/. The rapid transition from one to the other is perhaps analogous to the
semantic theme in a way that could provide a basis for the association.

On the other hand it can be argued that the individual phonemes are the
source of the theme through the semantic factors that they are associated with, /l/
with its factoral meaning display, conspicuous visibility, and /k/ physical intensity
(which can be expressed in many ways, as we will see). This is most obvious in the
words that describe sounds in the second group above. The phoneme /l/ appears to
operate as it did with the bl- words in the word under 4 in which loudness, a form of
display, is a common feature. But this is clearly not the case with the words under 3. I
suggest that tight grip seems to have a sub-theme, a diametrically opposed notion,
the exclusion of the light implicit in display. This is an effect of a tight grip. It is
consistent with an important aspect of the semantic factors that I will outline later –
their inherent polarity, display is integrally associated with concealment. This may be
an explanation of the polar thematic effect in, for example, clam, clasp, clench, cling
and so on – a grip so tight that it excludes all light. Interestingly some of the words I
have classed as not phonaesthetic (5), claim, clean, clear also have meanings that
involve the /l/ factor.

But what is the situation with the initial sound, /k/. Does it bear a meaning that
contributes to the meanings of the words? First, the sound. In English there are two
/k/ phonemes that are finely distinguished by the area on the roof of the mouth that
the tongue contacts in forming the sound. They can be clearly distinguished in come
and kick. In the first the tongue makes contact behind the palate (on the velum) while
in the second it is much further forward, on the palate. The difference is clearly
audible. In English orthography this distinction is made by using the symbols c and k
respectively. K is used only when it precedes the front vowels /e/ and /i/ and the now
silent (also frontal) /n/. In the case of the cl- cluster, however the initial phoneme is
formed in the more frontal position than in cart as is necessitated by the frontality of
/l/. (The practical impossibility of combining the /k/ of cart with /l/ is easily illustrated
by trying to combine those sounds). This palatal /k/ has a meaning in the system that
I will describe in Section 6. It is physical intensity that can be expressed in a variety
of ways such as complexity, angularity and kinaesthetic activity. It is this k-meaning,
rather than the very different c-meaning that I will describe later, that appears to
operate in the cl- cluster. It takes the form of tightness and intensity in most of the
core phonaesthetic cl- words. It can be perceived easily in clamp, clasp, claw, climb
and with only a little more difficulty in clap, clash, click and very mildly, it seems, in
cloak, clothe and cluck.

As a consequence of this analysis we can see that the phonaesthetic
meaning, tight grip, is an amalgam of the meanings associated factorally with the
individual phonemes while it is at the same time motivated by the articulation of the
two consonants. No wonder it is so effective.

My third example is the gr- words. Their meanings have a ‘feeling' that is broadly
similar to that of the last group but it is distinctive and different as are the sounds
associated with it and, I propose, the meanings associated with the two sounds.

1. /gr/
2. deprivation, seeking.
3. grab, graft (bribe), grapple, grasp, grateful, great, greed, grim, grimace,
---grime, grin, grind, grip, gross, ground, group, grovel, grow, grudge.
4. grace, graft (v.), grain, grand, grant, gratify, gravity, greet, grief, grist, gristle,
---grit, gritty, grizzle, groan, groggy, grope, grotesque, grouch, growl, gruesome,
---gruff, grumble, grump, grungy, grunt.
5. grade, grave (n.), grass, green, grocer, groom, groove.

This is a remarkably rich and productive phonaestheme but it is complex because the
common element of meaning is not easy to discern in all of the words and there are
some obvious evident contradictions. Part of the common element, I suggest, is
intensity that is palpable in one form or another, to some degree, in all the core words
of 3 and 4. But this is a different kind of intensity from that manifested in cl- words. I
propose that the meaning of this phonaestheme is a combination of the meanings
associated with the /g/ and /r/ phonemes as described Sections 6 and 7 where /r/ is
designated as having the primary meaning intensity of effort (distinct from the k
intensity) but it can also operate simply as an intensifier, a marker of experiential
salience. On the other hand /g/ has the meaning abundance, availability, generosity
but ranging through to its polar opposite, seeking, deprivation, grasping. The g
meaning is quite complex here because it has three distinct forms. It ranges across
the scalar spectrum of this phonaestheme from its positive pole, abundance etc. in
grace, grain, grant, gratify, great, greet, gross, group, ground, grow, to its negative
pole in grab, grasp, greed, grope etc.

However, this leaves the majority of the words I have classified as
phonaesthetic unexplained. What warrants claiming grapple, gravel, grimace, grime,
grind, grit, grovel, grudge, grumble, gruff
as members of the phonaestheme? Neither
pole of the phonaestheme as I have described it is readily evident in these words but
there is a prominent physicality and intensity that I attribute to /k/. This phoneme is
close in articulatory terms to /g/ except for its lack of voice. Thus we have the
apparently odd situation that /g/ is substituted for /k/ on account of its added force but
retaining the intensity that is highly consistent with the meaning of this part of the
phonaestheme. There are thus three distinct parts to the gr- phonaestheme, one
contingent largely on the positive pole attributed to /g/, one on its negative pole and
one on a harsh physical intensity associated with /k/ but without the presence of
either pole of /g/.

My last full example is one of the numerous terminal cluster phonaesthemes in
English, the -umble words, bumble, crumble, fumble, grumble, humble, jumble,
mumble, rumble, stumble, tumble.
The semantic theme that is common to most of
these words is a mild sense of disparagement. In the background of each word is a
‘normal' word or concept (respectively speak, break, hold, complain, be proud,
ordered, speak, make a deep sound (v.), walk, fall
of which the symbolic word is
almost a caricature.

Why is this combination of sounds associated with this disparaging
connotation? The -mbl cluster might be thought to have a role but this is not
supported by other words that contain this sequence, for example: amble, bramble,
gamble, ramble, scramble
in which disparagement is absent. The phonaesthemic
meaning there is quite different, something like random action. These words are not
freighted with affective content. This is a far cry from the meanings of the -umble
words. The related nimble, thimble also have meanings with little in common with the
first group. The -mbl cluster does not appear to carry a specific meaning of its own.

Perhaps the /u/ sound is influencing the meaning. In fact there is ample
evidence that /u/ is commonly associated in English with a sense of the
derogatoriness, disparagement or more generally displacement, disorder, being out
of kilter
, a semantic factor that we have already encountered. In fact disparagement
is a sub-factor of /u/. Many other examples (such as the –ump and -uck words) can
be cited, especially in monosyllabic words that are descriptive of physical entities,
actions and qualities. The basis of the meaning very commonly associated with /u/ in
this phonaestheme is clearly the /u/ semantic factor.

4.2 Characteristics of the phonaesthemes.
Despite its prominence in the English lexicon phonaesthesia has been regarded as
an anomaly and a puzzle on several counts: it contravenes saussurean arbitrariness
according to which phonemes, and by extension sub-morphemic clusters of
phonemes, should not be associated with specific meanings; it involves parts of
words that are not morphemes but that carry meaning that is generally regarded as
the privilege of morphemes; and it has a number of characteristics that have not
previously been well explained. The purpose of this section is to discuss some of
these characteristics with a view to obtaining a better understanding of it on account
of its strong relevance to the rest of the paper. Only word-initial phonaesthemes are
considered here.

4.2.1 The phonaesthetic groups.
As we have seen, the composition of these subsections of the dictionary
varies substantially. In some cases the majority of the words are phonaesthetic;
in others only a relatively small proportion are; some appear to contain more
than one phonaestheme; and some appear confused as there is evidence
of multiple or contradictory phonaesthemes. A fuller treatment of phonaesthesia
would need to account for these matters. The co-occurrence of phonaesthetic
and non-phonaesthetic words in the sub-sections indicates that the clusters
of phonemes have no categorical binding with the meaning they have in the
relevant words. Furthermore the same meanings can occur in other words
independent of the phonaesthetic consonant cluster. The examples demonstrated
that there is generally a penumbra of words in a cluster group that do not conform
strictly to its theme. They represent extensions of the meaning into new fields. The
boundaries of the phonaesthemes are hard to determine.

4.2.2 The character of the phonaesthemic meanings.
These are a curious blend of specificity––we know intuitively what they
mean––and fuzziness-–they are quite difficult to pin down in a definition.
This is due in part to the indeterminate boundaries of the phonaesthemes
with their often substantial penumbras of uncertain cases and, in some
instances, the arguable existence of more that one phonaestheme associated
with one phoneme cluster––but generally with a unifying motif. The prime
cause of the lack of clarity is that the meaning represents a gestaltic facet or quality
of human experience that is spread across a range of types of experience. It is not
surprising that they are hard to define. This characteristic imprecision enables them
to achieve considerable semantic diversity, for example by analogical extension into
different semantic fields.

4.2.3 Physicality, embodiment.
The great majority of the phonaesthetic words are characterised by being
strongly physical and sensory. They have a strong gestaltic character and
are generally graphically descriptive of concrete actions, things and qualities.
In some cases (mainly word-final phonaesthemes) there is also a prominent
affective content. The phonaesthetic meanings can be described as profoundly
embodied in that the words relate predominantly to actions and things with a strong
connection with the bodily experience.

4.2.4 Position in lexical space.
The words in this class are all content, as opposed to grammatical function
words. Verbs greatly outnumber nouns and adjectives. A large proportion
of the phonaesthetic words do not belong to the core of the lexicon. (I use
this term for the very small proportion of words in any section of the dictionary that
are very commonly used and are central to everyday human experience. The
meanings of more complex words are generally dependent on these words.) Only a
few can be found in the first 3000 words of any frequency-of-use corpus. The reason
for this lies in these words' position in human experience where the actions and
things they denote are generally derivatives of those that are denoted by core words.
(As gleam, glitter, glisten are semantic derivatives of the concept shine.) However,
most clusters have a nucleus of core words that generally belong to an older stratum
of the language. These can be regarded as catalysts for the historical formation of
the phonaesthetic groups of words (e.g. stand and still in the st- group).

4.2.5 Productivity in generating new words.
Phonaesthesia has been remarkably successful in English in generating
new words in the zone in which it operates––words that denote physical
actions and entities with strong sensory connotations. In numerous cases
this productivity has operated, not only in the English language, but also in
its precursors, evidently back to their Indo-European roots. The productivity of
phonaesthesia is clear through the history of the English language and continues
at a modest rate.

4.2.6 Cognate words in phonaesthemes.
In many phonaesthemes there are groups of cognate words with etymologies that
share common Indo-European or Germanic roots. This reflects a propensity to form
new words from old by creating new variants of the phonaesthetic meaning to express
related facets of experience. All phonaesthemes, however, have a wider non-cognate base.
They are not just assemblages of words with cognate meanings.

4.2.7 The ‘other half'.
The initial part of these words is their common active ingredient, phonologically
and semantically. What about the other half that has the function of transforming
the cryptic consonant cluster and theme into full-blown words? Is it simply an
arbitrary complement to the initial cluster that serves to mark the meaning
of the full word? Or does the sound-meaning contagion spread to the other half?
There is evidence that it does, to a certain degree. There are many examples of
words with an initial phonaestheme that are completed by a terminal one, or
something resembling one like –ffle (snuffle), -tter (flitter), -bble (squabble), -
ggle, (struggle), -uck (pluck)
. On the other hand many phonaesthemes display a
pattern in which just the vowel following the cluster contributes significantly to the
meaning. For example the well-known size symbolism involving open and close
vowels standing for, respectively relative largeness and smallness is quite active in
phonaesthemes as is the sound symbolism often associated with u in English where
it conveys a sense of displacement, disorder and disparagement. /O/ is another
element in this vowel symbolic system. It often conveys a sense of openness or
spatiality
. These four vowel sound symbols are copiously represented in the
phonaesthemic words listed in Annex 2 in ways that add specific semantic features to
the phonaesthemes. A clear example is in the gl- and gr- groups where glare; glint,
glimmer; grip, grope, grow, growl
; and grudge, gruff, grungy, grunt exemplify the way
these vowels make distinctive contributions to the words' meanings. Other
phonaesthemes demonstrate that this pattern is significant.

4.2.8 Are the phonaestheme clusters' phonemes conventional
or are they semantically motivated?

Phonaesthesia clearly owes its existence to a natural tendency to link
phonological form with facets of word meaning in the generation of new words
but a central issue is what determines the phonemes in the clusters that are
associated with specific meanings. Is phonaesthesia simply a means to allow the
expansion of the lexicon in a particular lexical field using arbitrary phoneme clusters,
or is the selection of the phonemes motivated in some way? As we have seen there
are two motivational possibilities––that the articulatory form of a cluster as a whole
has a character that ‘fits' the meaning, or that the individual phonemes contribute to
the meaning. It has commonly been assumed that the association in phonaesthemes
is predominantly conventional. This means that at some point in time an arbitrary
cluster of phonemes occurred in a core word of a phonaestheme and over time other
concepts that contained a facet of meaning in common with that word adopted the
same cluster with sufficient differentiation of the form of the whole word to be lexically
viable. Thus slide, slip or slow could have begun a gradual chain reaction, by a
process of imitation, leading eventually to variants on the original phonaestheme
such as slack, slop and sly.

We have seen that there is evidence for both the motivational possibilities
mentioned above. For example, in the case of articulatory motivation, in /sl/ there is a
very easy transition between the phonemes that seems highly appropriate for this
phonaesthemic meaning while /cl/ and /cr/ have more difficult transitions that could
have come to be associated appropriately with meanings of a ‘difficultʼ nature. The
five triple consonant clusters /scr/, /spl/, /spr/, squ- and /str/ also lend themselves to
such an interpretation. They have a relatively lengthy articulation time and involve
somewhat difficult transitions that seem particularly appropriate for their meanings,
for example, scramble, scratch, scream, screw and straggle, stream (v.), stretch and
stroke. However, the bulk of the evidence suggests that most of the clusters are not
primarily motivated by the way they are articulated. I have foreshadowed another
possibility that may have deep implications. This is that the individual phonemes
carry semantic connections that combine to form the phonaesthemic meanings. This
is a clear pointer to the subject of the next two sections.

4.2.9 The cognitive status of phonaesthemes.
A question psychologists have posedis whether phonaesthemes are
‘psychologically real'. Do they affect the way we process a word that contains
this feature compared with the way we process other words? This does not
require that we are conscious of them as we use them but rather that they
perform a specific and detectible psychological function. Priming studies have
suggested, for example, that phonaesthemic words are responded to faster
than words with other linguistic features such as semantic relatedness.
(Bergen 2004.) The value of such studies is questionable because they may simply
be demonstrating that the graphic quality of phonaesthemic words makes them
recognisable more quickly than is the norm. What is incontestable about
phonaesthemes is that they convey a distinctive albeit variable constituent of a
word's semantic content. In some cases this constituent is particularly prominent in a
word's meaning even if it does not reach conscious awareness, but in others it is
buried in etymology and performs no psychological function. Generally speaking,
phonaesthemes display psychological reality par excellence.

4.3 Summary conclusion.
The foregoing description of a small but typical proportion of the phonaesthemes of
English has revealed several distinctive aspects each of which is highly relevant to
the description of the semantic factors in the next two sections.

1. Sub-morphemic meaning. A small set of types of morpheme (grammatically
defined components of the form of words), particularly the root concepts of words, is
universally held to be the prime bearers of meaning in language. However, the
phonaesthemes demonstrate that there are some important sub-morphemic
bearers of meaning.

2. Non-arbitrariness. The phonaesthemes are a paradigm of non-arbitrariness in that
they are almost all motivated by either an iconicity involving cluster articulation and
thematic meaning or their composition from semantic factors.

3. Semantic generativity. We have seen in the examples that the central theme of
phonaesthemes diversifies into variant sub-themes that demonstrate their ability to
generate a range of lexical items according to some fairly well-defined principles.

4. Distinctive somatic character. The phonaesthemes all have a distinctive character,
a graphic physicality that distinguishes phonaesthetic words from most other words in
the English lexicon. Each theme has an idiosyncratic gestalt that is ‘felt' in a bodily
manner.

5. Dispensibility. Although the phonaesthemes have a significant position in
contributing to the English lexicon they are not necessary – they have no place in the
lexicons of many languages (such as Maori). They are not essential features of
language.

6. Psychological reality. There is a robust psychological reality in most phonaesthetic
words although in some it becomes buried in their etymology.

Phonaesthesia is a more familiar example of the ‘strangeness' of the unrecognized
semantic factors that will be encountered in next thee sections.


annotated contents




5. A new form of sound symbolism: articulatory iconicity.

5.1 The characteristics of the semantic factors.
This section introduces the central sections of this essay, the next two, that describe
in considerable detail the 24 semantic factors as they are represented in the initial
sounds of core words in the English lexicon. The purpose of these sections is to
establish firmly the actuality and validity of these new linguistic entities and their
significance in the composition of word meaning.

Although the semantic factors have a lot in common with the phonaesthemes as we
have just seen they are starkly different in several respects:

.........1. Although the factors were discovered through an association with
............. specific sounds, they are not dependent on this association and occur
............. far more commonly throughout the lexicon without it.
......... 2. The meanings of the factors are simple, irreducible and much more
............. basic elements of word meaning and human experience than the
............. phonesthemes.
......... 3. The factors have an entirely different nature as they are much more
............. abstract although marked by an idiosyncratic gestaltic quality similar to
............. that of phonaesthemes.
......... 4. Unlike the phonaesthemes they combine freely to form configurations
.............that provide the full structure of word meanings.

Researchers working with semantic features have generally assumed that they
should ideally have a number of specific characteristics to enable them to operate in
forming word meaning. The semantic factors possess all those characteristics, and
much more. They are, as we will see:

.........• few in number (24 in my estimation – see Table 1);
.........• highly abstract and very general in scope;
.........• both necessary and sufficient for creating the structure (rather than full
........... word meanings) of the meanings of the core words;
.........• arguably innate and universal;
.........• systematic in that they combine freely to produce lexical semantic
........... structures;
.........• parsimonious in their capacity to provide configurations of factors that
...........spell out word meaning at a structural level.

To anyone familiar with the literature, it will be clear that the existence of features of
this kind, if they are proven to be valid are a very surprising discovery that is likely to
impinge in interesting ways on a number of contentious matters in linguistics,
psychology and philosophy of language. But the factors may seem disappointing in
one key respect – like the phonaesthemes they do not generate the full meanings of
words. Additional semantic elements are required. Investigators in this field have long
been aware of this issue that has been dubbed ‘the problem of completers'
(Jackendoff 2002). This turns out to have a satisfactory solution (see Section 9).

Given the above qualities, the factors exhibit unparalleled, almost undreamt-of
capacities that are relevant to our understanding of language:

.........• they reveal lexical semantic content to an extent, and at a degree of
............ specificity, that has not been achieved before;
.........• they are capable of building the structure of the meaning of all types of
............ words in any language;
.........• under a new model of word meaning that recognises the central role of
............ the factors and supplements them with two other semantic elements
............ they can account for the full meanings of all words (Section 9);
.........• strikingly, this capacity applies to abstract and semi-abstract words and
............ the closed-class grammatical words that are resistant to other attempts
............ to describe their semantic content;
.........• they have connections with somatic regions of the brain that provide the
............ means for the mind to represent entities in the world;
.........• the factors may be relevant to the combinatorial capacity that is
............inherent in words' ability to form phrases and sentences making them
............ central to syntactical and grammatical theory; and
.........• they may consequently prove valuable to neuroscientists seeking to
............understand how word meaning operates in the brain.

If these capacities can be validated, the semantic factors will prove far more powerful
than any other set of semantic features that has been proposed. There are two
reasons why the factors have this unparalleled potential. First, they are not purely
linguistic because they also operate as fundamental constituents of human cognition
of the world, experience and behaviour
. Second, a large part of their remarkable
power stems from their division into the two main complementary types. The first can
delineate the physical structure of entities at a very abstract level and the second the
affective aspects of human cognition that underpin behaviour. It is also relevant to
note that the factors divaricate, like the phonaesthemes, into a number of sub-factors,
natural corollaries that facilitate their fit with the meanings of a vast range of words.

5.2 The semantic factors defined.
A semantic factor is a distinct, abstract constituent of word meaning, a constituent of
its internal structure, a ‘letter' of the ‘alphabet of meaning' that is essential for the
meanings of the words in which it occurs. The factors represent single, frequently
occurring facets of word meaning that are non-linguistic aspects of human cognition
that have a very high degree of primality, of being fundamental elements of human
experience and cognition. As such they must be common to word meaning in any
human language.

For centuries numerous thinkers (such as the English philosopher John Locke
who described word meanings as ‘bundles of ideas') have assumed that word
meaning was formed from configurations of sub-lexical semantic entities. They have
intuited the existence of such features but have been unsuccessful in identifying them
with any certainty up to the present time. In the last half century there has been
renewed interest by linguists but, despite a number of proposals, little agreement has
been reached (see Section 12). In the light of this history of failure my proposal faces
the enormous challenge of justifying the identification of something so elusive. But it
does so with confidence that is drawn from the quality if the evidence that is
presented here.

5.3 How the factors were identified.
The process of discovery of the factors was quite simple. It began well over a decade
ago with an almost casual observation of one section of the English dictionary that
provided the initial clue. I noted, as others must have done, that in the b section of
the English dictionary there is a substantial cluster of words that relate to the human
body and particularly its smoothly rounded parts – back, baby, belly, body, bosom,
bottom, breast, bum, bust, buttocks
and others. Words such as bald, bare and bathe
have a similar connotation. There is a comparable list of familiar non-bodily things
that are also distinguished by their convex form: bag, ball, balloon, basin, basket,
bath, bead, bean, beetle, bell, bird, bladder, bomb, bottle, bowl, bubble
and so on.
These words seem to have a common semantic feature that can be described as
bodily roundedness for the first group and simply convex form for the second. In
English these have become associated with the phoneme /b/ as the initial sound in a
substantial cluster of words. In the context of the large number of other b words that
do not have any hint of such a feature this may seem a mere curiosity. On the other
hand, if we extract the core words of this section (those that are most frequently used
and that have a particularly close association with fundamental human experience)
(say 40 or 50 words), this cluster of words begins to have a higher profile. However
there are other equally important clusters of b words with meanings of an entirely
different character, making it one of the more difficult sections in which to derive a
semantic factor.

As this result seemed interesting I investigated further with consequences that
surprised me greatly. I subsequently examined all the other sections of the English
dictionary by extracting small lists of core words (10 to 50 depending on the size of
the section) and investigating whether there was an evident semantic commonality of
meaning among them. The eventual result of this simple exercise was astonishing. In
most sections an abstract facet of meaning was readily apparent in an overwhelming
proportion of the core words. In others the situation was more complex but it was
possible to identify further facets once the pattern that was common across other
sections became evident. There was roughly one semantic factor in each section of
the English dictionary except x.

The result was the 24 semantic factors set out in Table 1. These all have the
qualities described above. It soon became evident that the basis of the identification
was arguably the association of the manner of articulation of the sound (or group of
phonemes associated with the initial letters of the words of each section, such as /s/
and /sh) with a single semantic factor. This association, articulatory iconicity, is
described in 5.5.

One characteristic of the factors that emerged early on facilitated the process.
In some sections the situation seemed to be confused by a significant proportion of
words with the opposite, polar tenor of meaning to that of the factor I had identified.
For example the factor energy or effort evident in the r section in such words as
rabid, rage, raid, rampant, rant, rape has a pole effortlessness, rest as in rain (v.),
reflect, remain, repose, remember, resemble and a scale of intermediate values in
other words. In some instances the two poles operate in the words of one section of
the dictionary but in others they are spread across two sections. I eventually
recognized that there was good reason to regard the factors as having a
fundamentally polar nature. Parameters of the kind that the factors represent are
essentially polar and scalar, although the poles are difficult to recognize in a couple
of instances.

The identification of the 24 factors was the beginning of a process of working
out their significance and their function in language, but one thing that was readily
apparent was that the presence of the factors was eminently verifiable and
empirically warranted. Although they are identified through an intuitive process rather
than direct observation, the process can easily be repeated by anyone. Once they
are recognized their presence in words is readily evident. Differences in the
interpretation of the precise scope of the factors or their presence in individual words
is inevitable, but the central picture is generally clear. The semantic factors appeared
to be something linguists had long sought, an empirically principled and definitive set
of constituents of word meaning.


..... Table 1. The semantic factors (abbreviated form).

1.
Materiality; magnitude (m) ...... 2. Positiveness/negativeness (d)
Particularity, specificity (p) Fullness/emptiness (v, z)
Surface, superficiality (f) Body and affects (s)
Bodily roundness, natural being (b) Tactility, deixis (t)
Contraction, compression (n) Abundance, generosity (g)
Extension, stretching out (y) Possession, self-interest (h)
Display, manifestness (l) Personal energy, humour, (j)
Openness, spatiality (o) Uncertainty (q)
Action; largeness (a) Displacement, disorder (u)
Smallness; interiority, identity (i) Existential value (w)
Intensity of energy (r) 3. The community life-world (c)
Physical intensity (k) The physical life-world (e)

The factors are described more fully in Annex 1.

The factors are described more fully in Annex 1.

The terms I have used in Table 1 to describe the meanings of the factors are
the best I have been able to formulate in some years of working with the factors but
this is still a work in progress at a very early stage of development. Differences of
opinion are unavoidable and should be fruitful. The work urgently needs the
involvement of other researchers. It will take some time and persistence for any
reader to become familiar with the results of my investigations because the factors
involve a new form of analysis of word meaning.

5.4 Why were the factors not identified earlier?
Given the simplicity of the discovery process it is rather surprising that they were not
discovered long ago. One might have expected earlier linguists such as Jespersen,
Firth and Jacobson, who had some sympathy with the sound-meaning relationship of
words, to have made a start, but that did not happen. Under the aegis of Generative
Grammar there was a flurry of activity in the 1960s and 1970s in endeavouring to
rehabilitate semantics by investigating the composition of word meaning (Katz and
Fodor (1963), Weinrich (1963, 1966), Bierswich, 1971). The postulation of semantic
features was an important part of this activity but investigators eventually concluded
that the task was too difficult and most turned their attention to other matters. Fodor
(1975) famously came to the opposite conclusion, that lexical meanings are pure
denotations that are atomic (featureless) and innate givens without internal content.

But why did no one follow the numerous hints in the English dictionary, such
as the body-related b words? The cognitive linguist Ronald Langacker, in an early
book (1967) noted the incidence in English of large numbers of words with medial
vowels u and i in words that have distinct common facets of meaning that he
described as, with respect to u, heaviness, dullness and filth and a generally negative
connotation; and in i, rapidity and insignificance. He went on to say that ‘we should
be careful not to overstate the importance of this kind of sound symbolism…..Sound
symbolism is not imaginary, but neither is it very powerful'. Langacker had had a
‘close encounter', but he failed to realize the significance of what he saw or to
investigate it further. An opportunity had been lost. In fact he had come close to
identifying two important semantic factors, displacement, wrongness, being out of
kilter
and smallness.

Langacker regarded any more substantial association of word sound and
meaning as being abjured by the Saussurean principle of the overwhelming
arbitrariness of this relation – as would any of his colleagues who had similar
encounters. Saussure's arbitrariness is a principle with an iron grip on linguistic
thought. It has been virtually unchallenged in the past century. Sound symbolism has
become an insignificant backwater of linguistics. The claimed discovery of a new
property of language through sound symbolism will be of much surprise to linguists.
Perhaps this discovery could only have been made by an outsider.

5.5 Articulatory iconicity.
Articulatory iconicity, the use of the vocal parts of the mouth to form shapes and
actions that mime prominent constituents of words' meanings, is a basic principle of
what I am describing. Sound mimes meaning (Nanny and Fischer 1999) in significant
ways that are most obvious in size symbolism (the use of back and front vowels in
words that denote size as in large and little, far and near) which is common in many
languages, but articulatory iconicity is endemic in English and some other languages,
as we will see. It is an eminently natural way of expressing aspects of the meanings
of words but there are severe practical limitations on the extent to which it can be
utilized within a word. It is a surprise, however, that it has a powerful capacity to
express such abstract facets of word meaning, the semantic factors.

The iconicity that is involved here is in the form of an association of (i) facets
of word meaning that contribute to the full meaning of words and (ii) sounds with
specific articulatory characteristics that enable them to form an iconic relationship
with afacet of meaning. Iconic articulatory gestures provide important clues to the
meanings of words, but not their full meanings. They occur most commonly in the first
sound of English words, but also in other positions to a lesser degree. The factors
themselves do not just occur singly but in combinations or configurations that spell
out the structure of the meaning of many core and near-core words. Far more often
than not, the factors occur without any iconic association with a sound, simply as
semantic features. The factors are found most saliently in the core words that are the
historical bedrock of the language. These are the words most closely associated with
ordinary human experience, the so-called basic-level words. In these words they
often spell out the meaning to a considerable degree as we saw with ‘place' earlier.
Most of the core words of English are of Germanic origin and can be traced back to
ancient Proto-Indo-European roots although many are from Latin and French. This
iconicity is generally much less evident in the far larger number of non-core words
that have entered the language more recently or have been fabricated in some way
in historical times. I need to emphasise again that the iconic representations of the
semantic factors are the tip of the iceberg of the factors' presence in word meaning
as they occur far more frequently without any association with the sounds of words.

I will explain how this articulatory iconicity operates with specific sounds in the
next two sections as I work through the sections of the English dictionary. The
question as to why it occurs is hard to answer but a plausible suggestion is that it is a
natural way to marry meaning and sound in words and also a valuable aid to learning
and understanding words and to communication as well. This is more obvious in
hand gestures of sign languages in which the prominence of iconicity is widely
acknowledged, and where the advantage over arbitrary signs is clearer. We are
rarely consciously aware of this vocal iconicity in our words.

I have claimed that the semantic factors are universal elements of language.
This may seem unwarranted because my main investigation has been in English. My
claim is founded primarily on the factors that emerge from English, but more limited
investigations have confirmed that the same iconic principle operates in some other
European languages, although somewhat less prominently than in English. I have
examined only two non-Indo-European languages in any detail, Finnish and Maori
(Lloyd, 2013). Articulatory iconicity is surprisingly prominent in both.


annotated contents




6. The meanings of vocal sounds: the descriptive factors.

In this section and the next I will work through, one by one, the full set of semantic
factors as I have identified them. It should become increasingly clear how the initial
phoneme/sound of many frequently used words in each section of the dictionary can
be seen, with a generally high degree of confidence, to be associated with a specific
facet of word meaning. As we work through the sections the remarkable nature of this
phenomenon will become apparent.

The full set of semantic factors is set out in abbreviated form in Table 1 above.
Fuller descriptions are provided in Annex 1.

The order of the sections as I present them is different from that of the
standard alphabet that has no specific rationale. The order I have chosen has its
logic. The sections of the dictionary are examined in three parts to distinguish three
distinct kinds of facets of meaning that emerge. One is descriptive of abstract
physical features of the world such as surface and action and the other represents
something totally different, stances of affectivity or ‘gut' feeling that are facets of the
meaning of many words to a greater degree than is normally recognised. The third
are factors with more generic taxonomic functions. The order I have adopted for the
alphabet of meaning is:

........1... m, p, f, b, n, y, l, o, a, i, r, k
........2. ..d, v/z, s, t, g, h, j, q, u, w
........3. ..c, e

The basis of the investigation in this section and the next is word-lists drawn
from the sections of the English dictionary. These lists contain all the core words as I
have identified them plus some other frequently used near-core words to provide
further illustrative material. The selection of core words is based on my own
judgement of ‘coreness'. I have tried to be objective but there is plenty of scope for
disagreement about the inclusion and exclusion of words. We will see that the pattern
that emerges is insensitive to variations at the margin.

A much more significant intuitive judgment imposed on the words is whether or
not they incorporate the semantic factor that I propose is present in the core words of
a section. The process might seem to have the potential to become impossibly
bogged down in disagreement but I believe the reader will find that it produces
satisfactory results despite inevitable differences on points of detail. This is an area in
which there is a substantial and irreducible degree of fuzziness in the data but the
results that emerge are clear and immensely interesting in terms of their implications.

In the lists below the core words that I have judged to be iconic are in bold.
The remainder are near-core words that are weaker in terms of the determinants of
‘coreness', such as frequency of use, but they provide useful examples. Those that
are asterisked are polar words, those with a meaning that is the conceptual opposite
of the factor description heading. These are integral parts of the factors on the
principle that the head factor necessarily implies the polar meaning and some
graduation between the two poles, as ‘hot' implies ‘cold' and the intermediate range
of temperatures. In some cases the poles are represented by other sounds and other
sections of the dictionary.

Following the word lists the sub-factors of each factor are set out. These are
natural ‘corollaries' of the head factor that represent ways in which the factor
meaning extends along natural lines that express closely allied areas of meaning of
which polarity is one. The sub-factors are normally mandated by a group of words
from the section.

6.1. The meaning of m: materiality
The facet of meaning that I have identified as most prominent in this section's core
words is materiality, an ontological feature of the referents of these words and many
others. A selection of iconic m words follows. Core words are in bold and those with
polar meaning are asterisked. Non-iconic words of the same core and near-core
status are also identified. The comments that follow cover (i) any interesting features
of the factor and the section; (ii) an explanation of how the articulatory iconicity is
motivated; and (iii) comments on polarity.

Machine, magic*, magnificent, main, major, make, male, man, manner*,
manufacture, manure, many, mark (n. and v.), mash, mass, massage, mat,
material, matter, me, meal, mean* (v.), meaning*, measure, meat, mechanical,
medicine, medium (a.), melody*, melt, mend, mental*, mercy*, mere, merge, merit*,
mesh, message, metal, meter, middle, milk, mile, mill (v.), mince (v.), mind*, mine
(n.), mineral, minimum, minor, mint (v.), minus, minute (a.), minute* (n.), miracle,
mix, moist, moment*, money, month, mood, moon, moral, more, mould (v.),
mould (n.), mound, mountain, move, much, muck, mud, muddle, murder, muscle,
music*, mutton, mystery*, myth*.

Non-iconic words: mad, mail, market, marry, may, mean (a.), meet, member,
metaphor, method, mild, mirror, miss, mist, mistake, mock, moan, modern, mourn,
mouth, moth, moss, murmur, must, my.

There are seven sub-factors. Here and in the following sections the polar meaning
is marked with an asterisk

........1. Materiality, substance: manure, metal, milk, mineral, mountain, muck, mud.
........2. Association with materiality: make, measure, meet, melt, money, moon.
........3. Extendedness, horizontality: massage, mat, more, most, much.
........4. Magnitude, multiplicity: magnificent, many, mass, minor, more, much.
........5. Measurement: main, measure, medium (a.), meter, middle, minute.
........6. Manipulation of materials, artifactuality: make, melt, merge, mill (v.), mix,
............ mould (v.).
........7. *Immateriality, the mental: magic, meaning, mind, mood, music, mystery.

Content. This range of sub-factors makes this quite a complex factor. It is robust in
terms of its representation among the core words with some 70 percent of the core
words being iconic under the interpretations that I have made. By definition
materiality is a feature of concrete nouns in all sections of the dictionary but this
section is unusually rich in words in which materiality is particularly salient. The factor
has the important function of distinguishing one of the two classes of nouns, mass as
opposed to count nouns. The first cannot take a plural while the second does (milk
compared with mark).

Motivation. Can we identify anything in the articulation of /m/ that might motivate this
particular association of sound and meaning? In phonetic terms this sound is defined
by five phonological features as a frontal, voiced, nasal, bilabial, continuant
consonant. Nasality and bilabiality themselves do not have any apparent relevance to
the iconicity. But continuance does. An essential feature of materiality as a physical
ontological feature is that it pervades whatever it is present in, as it pervades the
physical world. This is particularly obvious in the mass nouns, meat, metal, milk and
so on. This contrasts sharply with the plosive quality of /p/ and its very different
factor, particularity. Two other articulatory features are also relevant, voice and
frontality. The presence of voice can be seen to have an analogical relevance in
relation to the solidity of most material entities. Finally, a surprising aspect of the full
set of factors and associated articulations is that place of articulation is often relevant
to the choice of semantic factors that are associated with specific sounds. In brief,
frontal consonantal sounds are generally associated with descriptive factors that
specify physical features of entities. This applies to all the consonants with
articulation that involves the lips, except /v/ and /w/. In contrast, sounds that are
articulated well back in the articulatory space are generally associated with the other
main type of semantic factor, the affective factors. We will see that /g/ and /h/ are
prime examples of this type. When the full articulatory/factoral system unfolds this
explanation should become more credible than it may at the moment.

Polarity. The polarity of immateriality is robust in terms of its representation in words
in the section. It is a natural, integral feature but we will see shortly that the m factor
is unusual in that it seems to have a dual pole, one being in the next section of the
dictionary.

The m factor illustrates the character of the whole set, especially their abstract
specificity in representing very fundamental aspects of meaning and cognition,
aspects that have the appearance of being necessary, universal and innate.
Materiality, as one of the most basic of the factors warrants its position at the
beginning of the alphabet of word meaning. It is so basic that it is simply taken for
granted in many of the words in which it occurs, but its presence in its substantive or
polar form is arguably the logical starting point for the construction of the meaning of
a word.

6.2. The meaning of p: particularity.
This very large section of the English dictionary provides an even more robust
example of the alphabet of word meaning. Here is a set of core and near-core words
in the same format as in the last section.

Pace, pack (v. and n.), page, pain, paint (v.), pair, panel, parcel, parent, part,
particular, party, pass (v.), passion, patch, path, patient (n.), pattern, pause, pay,
pea, peak, peck, peculiar, peel, pencil, peep, (v.), penetrate, perch, perfect, perform,
period, perish, permit (v.), person, persuade, pet, phrase, pick, picture, piece,
pierce, pile, pill, pillar, pin, pinch, pip, pipe, pit, pity, place, plain (n), plan (v. and n.),
plane, plant (v. and n.), plate, please, plus, pocket, point (v. and n.), pole, pool, pop,
port, position, possess, possible*, post (n.), pot, power, practice, praise, pray,
prefer, present (n., v. and a.), press, prevent, prey, price, prick, pride, prince,
principal, principle, print, private, privilege, prize, probable, problem, process,
produce, profit, progress, project (v.), promise, prompt, pronounce, proper, property,
protect, protest, proud, prove, provide, pry, public*, pull, pulse, pump, punch,
punish, pure, purpose, pursue, push, put, puzzle.

Non-iconic words: paint (n.), pale, paper, patient (a.), past, physical, pig, pink, plain
(a.), play, plough, poor, popular, pretty,

This is one of the most straightforward factors. The unique semantic character of the
core words is evident almost at a glance. The head factor that I identify is
particularity, discreteness, specificity, the sense of being singular, bounded and
self-contained. While particularity is the central sense of this factor it has a number of
sub-factors that are all represented in the core and near-core words:

........1. Particularity, discreteness: pack, parent, path, pause, piece, put.
........2. Exclusivity: perfect, price, pride, private, promise, property, pure.
........3 (i). Seriality, severality: part, pace, page, pillar, pulse, punch.
........3 (ii). Verticality: pile, pillar, pole, post.
........3 (iv). Long and thin: pencil, pin, pipe, pole.
........3 (v). Flat and bounded: page, panel, patch, plain (n.), plank, plate, pool.
........3 (vi). Small and compact: parcel, pea, peep, pip, prick.
........4. *Generality, multiplicity: people, possible, public.

These move away somewhat from the head factor by being less abstract.

Content. The section is notable for the very small number of core words that do not
conform to the factor to a significant degree. Given my interpretations, around 90
percent of the core words have an identifiable p-iconicity. This is another factor that is
ubiquitous to an unusual degree in the lexicon because it applies (mainly without any
sound association) to all count nouns and to verbs that denote finite acts, telic verbs.
Telicity occurs in verbs that relate to a complete action, one with a goal, pass as
compared with please, an atelic verb as it has no specific end-point. A significant
number of p words are paradigms of the factor, for example: part, path, pause,
person, place, position, push, put
.

Motivation. The motivation of the association of sound and meaning in this section is
one of the most perspicuous. /P/ is characterized phonetically as an unvoiced bilabial
plosive. Its brevity and lack of the weight that voicing can impart seem to be
consistent with this semantic feature that is one of the most abstract that we will
encounter.

Polarity. A peculiar feature of this section is that polarity seems to be almost nonexistent
in the p-words. The logical pole of particularity is generality, multiplicity but
only three core words seem to bear this meaning. The factor also has a natural
polarity with m that is powerfully evidenced in the important count/mass noun
distinction. This is an unusual situation but we will find that it is replicated in the f and
b sections.

3. The meaning of /f/: surface.
This section also is prolific in its production of iconic words, but this and several other
of the twelve descriptive factors (b, n, y, a, i) can be seen to be more focussed than
the first two because the ontological features that they represent are more specific
and less ubiquitous than materiality and particularity and are of a distinctly different
character. They represent formal, topological aspects of physical entities. (This
applies here to the head factor and the second sub-factor but not to the others). The
semantic factor that I have identified, surface, appears in a variety of ways in the
world, in word meanings and the sub-factors, some of which may seem to be
somewhat obliquely connected with the head factor.

Face, fact, fade, fail*, faint (a., v.), faith, fake*, false*, familiar, family, fan (n., v.),
fancy*, farce*, farm, fascinate, fashion (n., v.), fasten, fat, father, fault*, favour, fear
(v. and n.), feast, feather, feature, feeble*, feed (v.), feel, feeling, feign*, fellow, felt
(n.), female, fence, fern, festival, fetch, fever, fiction*, field, fight (v.), figure, film,
filth, fin, find, fine, finger, fire, firm, fit (v.), fix, flag, flamboyant, flame, flank, flap,
flare, flash, flat, flatter*, flaunt*, flavour, flaw*, flay, fleck, flee, fleece, flesh, flexible,
fling, flip, flirt, flit, float, flog, flood, floor, flop, flourish, flow, flower, fluff*, flush*,
flutter, flux, fly (v.), foam, fog*, foil (n.), fold, folk, follow, folly*, fond, food, fool*,
foot, force, forest, forge, form (v. and n.), forward, foul, foundation, fragile*, frail*,
frame, fraud*, freckle, free, freeze, frenzy, fresh, friction, friend, frighten, frill*,
fringe*, fritter*, front, frost, froth*, fruit, frustrate*, fry, fuck, full, fumble*, fume, fun,
function, fundamental, funny, fur, furnish, fury, fuse (v.), fuzzy*.

Non-iconic words: factor, fair, fall, fame, far, fast, fatal, fate, few, fibre, final, finish
(v.), fish, first, fist, five, flea, fly (n.), focus, for, forbid, forget, forgive, fortune, four,
fraction, frequent, frock, frog, from, fuel, funeral, future.

A fuller description of the factor follows:

........1. Surface: face, fat (a.), feature, feel, figure, firm, flesh, form, front.
........2. Flatness: field, film, flag, flank, flap, flat, floor, flood, flow.
........3. The manipulation of surfaces: fashion (v.), fasten, fit (v.), fix, fold, forge,
............ form (v.), fuse (v.).
........4. Physical and personal intimacy: family, father, fight, folk, friend, fun, fuck.
........5. *Excessive detail: fancy, fluff, frill, fringe, froth, fumble, fuzzy.
........6. *Insubstantiality, falseness: fail, fake, feign, flatter, flaw, folly, frail, fraud.

Content. The presence of the factor in some of the words in the list may not be
evident at first sight but I hope a little cogitation on their meanings will confirm their
conformity with one of the sub-factors. In fact the factor has a metaphorical presence.
A fact is necessarily associated with a degree of confidence that rests on something
analogous to a firm, reliable surface that is polar to false. Faith likewise involves a
firm surface-like quality upon which one can depend but sub-factor 6 is sometimes
also implicit, representing the element of doubt that inherently threatens faith. To be
fascinated is to be preoccupied with surfaces (that may be illusory). The noun fashion
has a strong association with the aesthetic appeal of surfaces and forms, alluring or
just transitory, depending on one's perspective. Fear is a body-based feeling that
manifests itself in several ways including sensitivities of the skin. The iconicity of
feast, feed and food derive from the surfaces that are integral to these concepts
together with the materiality of m (as in meal and meat). Female is arguably f-iconic
(m and b are also prominent) on account of the variety of rich implications of surface
in the word's connotations. Finding cannot dispense with the visual and tactile
senses of surface. The scope and richness of the factor is outstanding. None of
these ‘explanations' is fully perspicuous partly because other semantic factors are
prominently involved in the meanings of all of these words. Some words, that I have
classed as non-iconic, can arguably be viewed as iconic. These include the core
words fall, fame, far, few, finish, for and from, all of which seem to have an obscure
element of surface in their meaning.

Motivation. What is the evidence of an articulatory motivation here? /F/ is formed by
the forced emission of breath between the lower lip and the front teeth without any
voice being produced. (The equivalent voiced sound is /v/ which, interestingly, has a
closely related factoral sense). In comparison with the other sounds that are formed
in the front of the mouth, p and m, f uses the surface of the lower lip in a distinctive
way that forms a credible analogue for the idea of surface. If there is an unconscious,
natural signalling system built into the way in which sounds are articulated in the
English language, as I propose, /f/ fits the bill to a high degree. This is a significant
phonological clue to word meaning.

Polarity. A question that arises here, as it did in connection with the m factor, is
whether polarity operates within the section or in relation to another factor. I have
concluded that the last two sub-factors, excessive detail and insubstantiality,
falseness are polar on the basis of a sharp contrast with the more positive aspects of
surface of the other sub-factors but there is another possibility. This is that there is
polarity between the surface of f and the bodily roundness of b that will be examined
next. I will try to resolve this matter at the end of that section.

4. The meaning of b: bodily roundedness.
This section generates another very large word-list that is perplexing in terms of
trying to identify a single semantic factor.

Baby, back (n.), bag, bake, bald, ball, balloon, band, bang*, bank, bar, bare, bark
(v.), barrel, base, bash* basin, bat, bath, bathe, batter*, battle*, bay, be, bead, beam,
bear (v.), beard, beast, beat*, beauty, bed, beg, bee, beetle, bell, belly, bend,
berry, big, bio-, bird, birth, bladder, blast*, block, blood, bloom, blossom, blow
(n.), board, boat, body, (v.) bold, bomb, bone, book, boom, boot, born, bosom,
botany, bottle, bottom, bow (v.), bow (n.). bowels, bowl, box, boy, brain, branch,
brawl, brawn, bread, break, breast, breed, bride, broad, brother, brow, bruise*,
brute, bubble, bucket, bud, build, bulb, bulk, bull, bum, bump (n., v.), bun, bundle,
burden, burst, bust, buttocks, bush, bust, butcher, butter, button, buxom.

Non-iconic: back (adv.), bad, balance, bark (tree), basket, beach, benefit, because,
become, before, behave, behind, behold, believe, belong, below, beside, best,
bereave, better, between, beyond, beware, bit, bite, bitter, black, blade, blame,
blank, blanket, blaze, bleak, bless, blind, blink, blow (v.), blue, boil, bond, border,
borrow, both, brew, brick, broom, bridge, bright, bring, brown, brush (v. n.), bunch,
bury, burn, business, busy, but, by.

A number of sub-factors can be identified here:

........1. Bodily roundedness, animality: baby, bare, beast, beauty, belly, big, body,
............ bone, boy, brain, breast, bum, bust (n.).
........2. Rounded objects and hollow containers: bag, ball, bead, bee, beetle, bell,
............ berry, bladder, bread, bug; barrel, basin, basket, bath, boat, bottle, bowl,
............ bucket.
........3. Natural kind, animacy: be/being, bear (v.), bio-, branch, brother.
........4. *Physical violence: bang, bash, battle, beat, bite, blast, blow (n.), bruise (v.).
........5. *Inanimacy: basin, ball, bath, blade, board, book, bubble.

Content. There are two important clusters of b words that I initially classified as iconic
but subsequently changed my mind. These are the prepositions: before, behind,
below, beside, between, by
, and the verbs: become, behave, believe, behold, belong.
Both sets have important roles in language but neither can be regarded as iconic in
the b sense. Their second syllable roots are however largely iconic with respect to
their initial phonemes.

The first two sub-factors represent much the same topological feature,
convexity. This is a typically biological form that results from the growth of animals
and organisms. The artifactual objects (bag, ball, bead etc.) were probably originally
derived from animal and natural shapes such as, in early times, bags made from
animal skins and rolled balls of clay. This biological theme can be taken to cover
other words in the section such as beak, bill, bio-, blood, bloom, blossom, branch,
bull, bush
. The penultimate sub-factor may be an anomaly that has been created by
a sound symbolic interpretation of /b/ as an analogue for violence. Consistent with
this is the several b expletives bastard!, blow!, blast!, bugger! It could alternatively be
regarded as the pole of the factor but it lacks the substance and true sense of polarity
for this to be taken seriously. There is a much higher proportion of non-iconic words
in this section.

In this and the previous sections one might have reasonably expected a
random distribution of these kinds of words through the dictionary as a consequence
of the operation of saussurean arbitrariness. This would yield very modest
proportions of words of any factoral type in any section. The results that the analyses
are producing are far out of line with such random distributions.

Motivation. What can be said about motivation in this instance? I propose that there
is a possible motivating factor. This is a recognizable similarity between the
articulation of /b/ and the sense of what seems to be the root of the factor, bodily
roundedness
. The production of this voiced plosive results in a slight increase in the
pressure inside the mouth with a consequent almost imperceptible puffing of the
cheeks, a feasible articulatory analogue for bodily roundness. The effect is easier felt
than seen. This may seem far-fetched but in the context of motivation in the full
factoral system it is quite credible, I believe.

Polarity. Finally, the question of polarity in this section, that I touched on above and in
the previous section. There are two contenders for a pole for this factor. One is that
the surface of f and that of b are in a polar relation. There is a polar-like tension
between the root sub-factors of f, the first two, and the first two sub-factors of b –
inert flatness and biological swelling. The second is that the violence cluster forms a
pole. This is unattractive because the cluster seems to be something of an anomaly. I
favour the first proposal. The other polar feature, inanimacy, is a more natural pole.
We have a messy situation similar to that between m and p.

4.8 The meaning of n: contraction.
This section is unusual on account of two sub-factors discussed below. N has a
semantic factor that I describe as: contraction, compression, reduction, focus.
Here are some illustrative words that demonstrate a very different form of articulation
and a very different semantic ‘feeling' among the words:

Nail, naked, name, nap, narrate, narrow, nasty, nation, native, nature (the nature
of), naught, navel, navigate, near, neat, necessary, neck, nectar, need, needle,
negative, neglect, neighbor, neither, negotiate, nerve, -ness, nest, net, never, new,
next, nexus, nibble, nice (good), nice (precise), nick, niggle, night, nil, nimble,
nipple, nitpick, no, noble, nod, node, noise*, nominal, nook, noon, normal*, north,
nose, not, notch, note, notice (v.), nothing, notion, notorious, nought, nourish, now,
nub, nucleus, nugget, numb, nut, nuzzle.

Non-iconic words: Nature (with a capital n), nephew, niece, nine, noun, nurse.

Sub-factors.

........1. Contraction, compression, focus, reduction: nail, naked, narrow, near, neck.
........2. Self-centred action, intransitivity. (See below).
........3. Negation: no, never, nothing, nought.
........4. *Dispersion, emission, release, expression: noise, normal.
........5. *Other-directed action, transitivity). (See below).

Content. The first perusal of the word list may not result in a clear impression of the
semantic factor in some instances. Contraction operates in many ways and in many
fields. Some elucidation is called for:

........Naked – reduction from a normal clothed state.
........Name – the exclusivity of nomenclature.
........Narrate, navigate, negotiate – following a tightly defined course.
........Nasty – a state resulting from the reduction of all goodness.
........Nation – the exclusivity of belonging.
........Navel – a focal point of the body.
........Neck – a narrow part of the body profile.
........Negative and its relatives – extreme reduction.
........• -ness – the essence of …
........New – without the concretions of age.
........Noon – a focal point of the day.
........Noun – the specificity of substantive denotation.
........Now – the ultimate reduction of time.
........Nut – the compression of a tree into a very small object.

Two unique factors, 2 and 5 have been included in this section. They do not occur in
an iconic form associated with this sound but they accord with the polar tenor of the
section. They represent intransitivity and transitivity and some other associated
general aspects of the verb function that are intrinsic and integral to the meanings of
verbs. These are universal constituents of the meaning of verbs that are often
attributed to syntax. The action of transitive verbs involves an outward-directed,
centrifugal feature that occurs, for example, in give, kick, kill, look, point. These verbs
possess two arguments, actor and patient (three in the case of give that requires
association with a complement). Other verbs, the intransitives, have the opposite
orientation in that they require only a subject, as with sleep, die, walk, fall, weep. This
distinction is a key feature of the sense of verbs that is highly consistent with the
meaning of this factor because it involves senses that can be described as
centripetal and centrifugal.

Motivation. The motivation for the association of articulation and factor is one of the
most convincing. This consonant is formed by the pressure of the tongue on the
alveolar ridge at the back of the upper front teeth. This pressure is an almost perfect
oral gesture to convey the factoral meaning analogically.

Polarity. The polarity of n is not evident at first sight. There are very few words in the
section that contain a polar sense. There is, however, another section of the
dictionary with a factor that can be regarded as polar to the meaning of n. This is the
next section, y, with its factoral sense of extension, stretching, the natural contrary of
contraction. It is significant that yes the indubitable pole of no falls within this section.
In the affirmation of yes we can sense extension in a metaphorical way.

4.9 The meaning of y: extension.
This section in the English dictionary is miniscule but it is another exemplar of
articulatory iconicity. Its core and other prominent words include: yard (measure),
yard (a domestic space), yarn (a tale), yarn (of fibre), yawn, year, yearn, yell, yes,
yesterday, yield (produce), yield (surrender), yoke (n.), yonder, you, young.

Non-iconic: yacht, yellow, yolk.

Y has only two sub-features:

........1. Extension, stretching, linearity: yard, yarn, yoke.
........2. Continuous, ongoing, duration, temporality: year, yesterday.

Content. The smallness of this section belies its importance in the full factoral
system. Extension is a topological feature that is very common in all sorts of entities
and their corresponding words while duration and temporality are existential
dimensions that must be represented in language. The inclusion of several words
might be questioned. I have judged yes as an iconic word on the strength of the
consistency of extension with the often vigorous affirmation that the word conveys.
You is perhaps more open to question. The relevant aspect is the contrast between
the first and second persons, I/we and you. You involves a reaching out, as it were,
to the other. In contrast the first person, I, we has a distinctive exclusivity and self-
centredness. The occurrence of the feature in young is plausible as extension here
implies the vitality of youth.

Motivation. is attributable to the peculiar form of the articulation of /y/ in which a glide
is formed between an unsounded /ee/ and the vowel that follows the /y/, miming the
factor very effectively.

Polarity. The pole clearly lies in the n section.

The meaning of /l/: display.

The l section of the dictionary is a fertile field for its semantic factor. The sub-features
are:

........1. Display, manifestness, conspicuous visibility: large, laugh, lay, leap, loud, lust.
........2. Illumination: lamp, light.
........3. *Concealment, obscurity, latency): lack, less, lock, lose, low.

Here is a list of core and other illustrative words in the section that represents one of
the most convincing examples of iconicity:

Lack*, laconic*, lady, lament, lamp, land, language, large, last (v., a.), laugh,
launch, law, lay, lazy*, lead (v.), leaf, lean (v.), leap, learn, leave, lecherous, leer,
lend, lens, less*, let, letter, libation, library, lick, lie* (v., position), lie* (v., n.
falsehood), life, lift, light (n.), light (a.), like (v., adv.), limbo*, limit, limp (v.), line,
linger, lip, list, listen, litany, literal, little, live (v., a.), livery, load, loathe, lock*,
locution, loose, lord, lose*, lost*, lot, loud, lour, love, low*, loyal, lucid, luck, lull*,
lullaby, luminous, lure*, lust, lyric.

Non-iconic words: lamb, left (a.), lemon, leg, lion, liver, lung.

Content. The l factor is present in many words throughout the dictionary but it is
especially salient here. While some of the iconic word meanings are perception-based
(lady, land, leaf) it also covers other modalities with equal facility: auditory
(laugh, loud), physical action (lean, leap), psychological acts (learn, lie, listen), size
(large, little), affectivity (like, love).

Motivation. The motivation is not immediately obvious. The sound's articulation
involves the release of the tongue from the alveolar ridge behind the upper front teeth
that, at the beginning of a word, seems to launch the following vowel and the rest of
the word into space. This is most evident in words with a matching intensity such as
large, laugh, leap, lift, loud, but it can also be felt to be present in most of the words
through a mutual reinforcement of iconic meaning and articulation. It is notable,
however, that the effect in the polar words is quite different as the iconicity loses it
force and the articulation seems relatively limp.

Polarity. The factor has a well-formed pole within the section.

8. The meaning of o: openness.
The factor that operates here has something in common with the l factor but is quite
distinctive. This may be obvious just from a reading of the word-list.

Oath, obey, object (n.), object (v.), oblige, oblivion*, obscene, obscure*, observe,
obstacle*, obtain, obtuse*, obverse*, obvious, occasion, occlude*, occult*, occupy,
occur, ocean, odd, odour, of, off, offend, offer, often, old, omen, omit*, on, once,
one, onerous, only, onward, ooze, open, operate, opinion, opportunity, oppose,
opposite, oppress*, optic, option, opulent, or, oral, orator, orbit, ordeal, order,
ordinary, organ, organize, orgy, orient, origin, original, ornament, ostentatious,
other, ought, our, out, outline, outside, outward, oval, ovation, over, overt, owe, own.

Non-iconic: oats, oil, onion, orange, orchestra, ore, ounce, oven, owl, ox, oyster

I have designated the semantic feature that emerges from these words as openness
with the following sub-factors:

........1. Spatial openness, spatiality: occupy, occur, of, off, on, open, out, over.
........2. Personal openness, mental space: observe, offer, opinion, ought, owe.
........3. Possibility, opportunity: opportunity, option, organize.
........4. *Concealment, confinement, obscurity: obscure, obstacle, occult.

Content. A distinguishing feature of the factor is the extension of the spatial into the
personal that is evident from oath to own. The section spans a remarkable variety of
words that represent psychological openness of various kinds – obey, oblige,
observe, offer, organize, ought
. This is one of the most robust iconicities. The kinship
with l is obvious. The sounds and meanings marry perfectly in the archaic lo! At the
pole is the Latin prefix ob-, which occurs in a variety of forms as the consonant is
modified by a following consonant. This prefix has several meanings but the main
ones that occur in the list are, against, hindering as in obliterate, obstacle (which
contribute to the factoral pole) and to, by as in offer, opportunity. The presence of
several closed class words in the section is notable.

Motivation. This is one of the most perspicuous instances of motivation with the
space in the oral cavity and perhaps the shape of the lips being excellent analogues
for the factor.

Polarity. The asterisked words constitute a well-formed pole.

9. The meanings of a: action, largeness.
The meanings that I have identified are, simply:

........1. Action, activity, process.
........2. Largeness, enlargement

These could be regarded as two separate factors but there is a strong connection
between them in that action generally has a fundamentally ‘enlarging' aim. It seeks,
basically, to promote the interest of the individual and the community in the face of
contrary forces. The substantial cluster of ‘enlarging' verbs in the section –
accomplish, accumulate, acquire, add, advance, affect (v.) – exemplifies this idea.
Here are the core and other illustrative words:

Abandon, able, abolish, about, above, absent, absolute, absorb, abstract, abundant,
accept
, accommodate, accomplish, according, account, accumulate, accuse,
acknowledge, acquire, across, act, active, adapt, add, admire, admit, adore,
advance, advantage, adventure, advise, affair, affect (v.), affect (n.), affection, afford,
after, again, agent, agitate, agony, agree, aid, aim, alarm, alert, alike, all, allocate,
allow, along, also, alter, always, amalgamate, ambiguous, amend, among,
apologize, appeal, appear, applaud, amount, amplify, amuse, analyze, ancestor,
ancient, and, anger, announce, annoy, another, answer, antique, anxious, any,
apart, apologize, appeal, appear, apply, approach, approve, arbitrary, architect,
area, argue
, arithmetic, army, arrange, arrive, art, ascend, ask, aspect, assemble,
assert, asset, assist, associate, assure, astonish, attach, attack, attempt, attract,
augment, authority, available, awake, aware, awful.

Non-iconic words: a/an, accident, accurate, acid, adult, against, age, air, am,
angel, angle, animal, ant, apple, arm, arrow, article, artificial, as, at, ash, atom,
audit, aught, autumn, axe.

Content. In all the vowel sections of the dictionary there are large clusters of Latinate
words with prefixes derived from Latin. These have a substantial influence on the
factor associated with the initial sound. The dominant prefix in this section, ad-, which
in Latin has a number of meanings including its spatial sense of to, toward, that has
the greatest influence here. (The prefix is often hard to recognize because the sound
of the consonant is modified in many words to a + another consonant to
accommodate the following sound). The function of the prefix in Latin was to turn a
noun or an adjective into a verb or to alter the nature of the activity denoted by a
verb. The resulting words are highly expressive of actions and include numerous
transitive as opposed to intransitive verbs. The prominence of these words in the
section contributes much to the action-oriented character of the section as a whole.
By contrast there is only one verb in the core n words. As I will explain later, I have
used a = action, activity as a marker for all verbs.

There is another prominent type of word in the section, some of the so-called
closed class words (so called because their number does not increase, unlike the
vast open class of all other words). No other section apart from b contains such an
impressive group of these words. Some of these bear a sense of enlargement,
augmentation, for example about, above, across, again, all, and that can arguably be
regarded as iconic. Others are clearly not.

The word-list illustrates once again the very strong iconic presence of the
factor in its words. Over 80% are iconic under my interpretations. The case for the
iconicity of /a/ does not rest solely on the words of this section. As I noted earlier, it is
widely recognized in intermediate positions as a form of size symbolism, a paradigm
of factor iconicity.

Motivation. As with the previous vowel the iconicity is based on the space in the oral
cavity. The distinction between the two iconicities is evidently the position of
articulation of the vowels, /a/ being further back. The relative frontness of /o/ seems
to be the deciding iconic distinction.

Polarity is found in the next section.

10. The meaning of i: smallness.
This section contrasts strongly with the last two in terms of articulation and meaning.

Word-list:
I, icon, id, idea, identity, idle, if, ignore, ill, imagine, impact, impediment,
imperceptible, imply, import, impression, impulse, in, incentive, inch, include,
incorporate, increment, indeed, index, indicate, individual, induce, indulge, inert,
infant, infect, infer, inferior, influence, ingenuity, ingredient, inhabit, inherent, inhibit,
initial, inject, inland, inner, innate, innovate, inquire, insect, insert, inside, insight,
insipid, insist, inspect, inspire, install, instance, instead, insulate, intake, integral,
integrate, intelligent, intend, intense, intent, intercept, interest, interfere, internal,
interpret, interrogate, interval, intimate, into, intrigue, intrinsic, introduce, intrude,
intuition, invade, invisible, invent, invest, inward, island, isolate, it, itch, item, itself.

Non-iconic words: ice, illuminate, illustrate, immediate, immortal, important,
impose, impossible, improbable, improve, inability, inanimate, inarticulate, incline,
incomplete, incorrect, increase, incredible, independent, industry, inefficient,
inevitable, infinite, inflate, infrequent, instruct, instrument, insufficient,
international, intervene, invite, involuntary, iron, irregular, irritable, issue.

The section has a diverse set of sub-factors that have substantial coherence and
iconic representation in the section.

........1. Smallness: inch, infant, inferior, insect, island, itch.
........2. Interiority: imagine, in, include, interfere, internal, into.
........3. Essentiality, identity: I, icon, inherent, innate, instance, intrinsic, it.
........4. Having an abstract or mental quality: idea, intellligent, inspire, intuition.

Content. This section is also heavily dominated by prefixed words drawn from Latin.
The Latin prefix in- with its primary meanings into, in and within has a heavy
influence. A second Latin in- prefix that means not, opposite of, without also occurs in
a considerable number of words such as inexpensive, inability, impossible that are
not iconic. The section is unique in the dominance of Latin-derived words and the
paucity of Germanic words. The two sounds of ‘in' in its various manifestations are
strongly reinforcing in factoral terms. The last two sub-factors provide the factor with
valuable scope.

The core iconic words constitute a smaller proportion of the total core words in
the section than in any other section. This is due in large measure to the numerous
in- = not words. This proportion does not in any way diminish the validity or the
significance of the semantic factor. The factor operates size symbolically far more
extensively in other positions in words where it is associated with smallness as a
factor in word meaning, often iconically. This is the polar side of the size symbolism
of /a/ and other open vowels.

Motivation. It is fairly obvious that the confined space in the mouth is the motivating
feature.

Polarity lies in the a section.

11. The meaning of /r/: intensity of action.
This section is a very different proposition.

Word-list.
Rabid, race (v.), race (n.), rage (v. and n.), raid (v. and n.), rain*, raise, rally,
ramble*, ransack, rant, rape, rapid, rapture, rare, ravage, rave, raw, re-, reach,
real, reap, rear (v.), reason (v.), reason (n.), rebel (v.), reckless, recline*, refer,
reflect, refuse, reject, rejoice, relate, relax*, relieve, rely*, remain*, remember,
remote*, remove, repeat, repel, repent, reply, report (v.), represent, require,
rescue, research, resemble*, resent, rest (v. and n.)*, retire*, retreat*, return,
reveal, revel, revenge, revere, rid, ride, ridicule, rife, rigid, rise, risk, riot, rip,
roam*, roar, rob, rocket, root, rot, rough, rouse, rout, row (n.), row (v.), rub, ruck,
rude, ruin (v.), rumour*, run, rush.

Non-iconic words: rat, rich, right (a.), right (n.), round, rust.

Sub-factors.

........1. Effort, energy: race, rage, rapid, rebel (v.), rally, ride, riot, rise, rout.
........2. Activity, movement, event, process: raise, reach, reap, rejoice, retreat, rise.
........3. Salience: rabid, rally, represent, resemble, right, rise, roar, rough, ruin.
........4. *Effortless, inactivity, rest, spontaneity: relax, remain, rest, retire, roam.

Content. The tone of the words is in sharp contrast to the sections of the dictionary
we have already studied. For example the first line of the l words that includes lack,
lady, lame, lament, lamp, land, language, large shows barely a hint of the intensity
we find here. In the r words it is possible to find a whole spectrum of the iconic
meaning from high to low intensity in words such as: rage, rape, ransack, rant,
reckless, revolt, roar, rapid, rough, rapture, rejoice, revere, rife, right, rule, remove,
reach, real, recognize, regret, remain, remember, resemble, resign, rest
. The words
become increasingly polar. Even /r/ gets tired! This semantic factor with its spectrum
of force is an important constituent of the meaning of verbs.

Motivation is not hard to identify here. /R/ has a very distinctive form of articulation
arising from the positioning and the tense shape of the tongue that give an
intrinsically energetic character that tends to be emphasized in words in which the
semantic feature is prominent. The articulation is a very appropriate analogue for the
semantic feature.

12. The meaning of /k/: physical intensity.
This is another of the small sections of the dictionary, but size does not diminish the
factor's application in word meaning through the dictionary.

Word-list.
Keel, keen, keep, ken, kernel, key, kick, kill, kind (a.), kind (n.), kindle, king, kink,
kiss, knack, knee, kneel, knell, knife, knight, knit, knob, knock, knoll, knot, know,
knuckle, knurl.

Non-iconic: kettle, kitten

Sub-features:

........1. Physical intensity, intensity of quality, distinctiveness, typicality: kill, kind (n.),
............ king, kiss, knight, knock, know.
........2. The prominence of parts, angularity: keel, key, kink, knee, knife, knob.
........3. The unusual form, complexity, refinement and intricateness of things: key,
............ kindle, knack, knit, knot
........4. (Generality, featurelessness). No exemplars in this section.

Content. This factor is manifested in various forms in the physical character of
entities and actions in many distinctive ways that make it quite distinct from the r
factor. At the level of abstraction at which the factors operate this factor can highlight
any kind of prominent idiosyncracy. It is a very useful component of the system in this
respect.

The interpretation of some of the words as iconic requires clarification. Kick,
kill and knock are clear exponents of the factor but in some of the words the intensity
is of a more subtle kind and implies that physicality includes physical aspects of
emotion and some forms of cognition. Words such as keen, keep involve an internal
tension of a proprioceptive kind. Key has a more oblique intensity that is located in
the intricacy and effectiveness of a key's interaction with a lock. Kind (a.) arguably
implies a controlled intensity in the application of compassion while kind (n.) implies a
differentiating distinctiveness. King (in the traditional historical role of this entity)
involves an intense concentration of power and prestige that is integral to the sense
of the word. Kiss has both physical and emotional intensity. In most of the kn- words
a somewhat different aspect of intensity is found. This is a distinctive angularity or
prominence of structural parts – in a knife, its sharp edge, in knee and knuckle their
physical angular prominence, in knit the intricate work of the flying needles and in
knot the tightly-bound complexity of overlapping cords. Finally, know seems an oddman-
out until the definitiveness of knowing, the precise fitting of concept and what it
denotes, and perhaps the difficulty of learning, are taken into account.

Motivation. The English /k/ sounds are represented by the letters c and k which
represent subtly different sounds. The /k/ of the present section occurs in English
words of early origin only before the e, i, and n (in which case it is no longer sounded).
As a consequence of its following sounds, /k/ is formed in the region of the hard
palate. By contrast the /k/ of the c section occurs before a, o, u, l and r and is formed
slightly further back in the velar region under the influence of these following sounds.
We will see later that the c section has an entirely different semantic factor. The
sharp factoral difference depends on this apparently miniscule difference in
articulation that is well-defined in articulatory and phonetic terms. This is a graphic
piece of evidence of articulatory iconicity. It is significant, however, that c words in
which the c is followed by /l/, namely phonaesthetic words such as clap, claw, clean,
clever
involve a palatal /k/ and have a distinct k (not c) semantic factoral character.
This also applies to cr- words even though the sound there is velar. The combination
of the consonants in this phonaestheme seems to transcend the wider principle.

There is a well-known psychological experiment that provides good evidence
of the k factor. This is the bouba/kiki phenomenon first noted by a gestalt
psychologist in the 1930s but more recently used by a prominent psychologist V. S.
Ramachandran (2001) as evidence of synaesthesia, the connection between
different sensory modalites in the brain. The experiment involves asking a group of
subjects which of the fictitious words corresponds to two displayed shapes, one very
spiky and one with bulbous curves. The overwhelming response is that bouba relates
to the smooth curved shape and kiki to the spiky shape. "This suggests that the
human brain is somehow able to extract abstract properties from the shapes and
sounds, for example, the property of jaggedness embodied in both the pointy drawing
and the harsh sound of kiki.” The basis for the interpretation of ‘bouba' is obviously
inspired by the b sound and factor (bodily roundedness). This factoral iconicity has
long been sitting under our noses without being recognized.

Polarity is not manifested in any of the words. I have specified the virtual pole purely
on the basis of the polarity principle. There is a logical need for such a pole but it is
interesting that it is also present in the pole the p section. This seems to be another
of the few untidy loose ends in the system.

This concludes the survey of the descriptive factors. There are only twelve of them
and they have an enormous task in constructing the physical/topological framework
of the meanings of words. Owing to the small number and the abstractness of these
factors, this lexical semantic framework is inevitably itself abstract, but it is often
surprisingly effective as later example will illustrate. In the great majority of words,
however, the descriptive factors require supplementation by the factors described in
the next section.


annotated contents




7. The meanings of vocal sound: the affective factors.

The affective factors introduce into word meaning emotive feeling in various forms
that have an extensive but under-recognized presence in language. In words like nail,
narrow, neck, page, part, pause, place
there is negligible intrinsic affectivity. But
many other core and near-core words contain affective constituents that are
fundamental to their sense. This can be a very obvious constituent as in nasty, nice,
no, noble, pain, pray, proud
but it is commonly present as a more subtle, deeply
embedded, existential element as in naked, name, nation, nature, near, necessary,
need, neighbor, nest, never, new
. We will see in the examples of the factoral analysis
of word meaning in Section 11 that the affective factors have a central role in word
meaning. They are a necessary and vital element of language because they
represent the psychological motivating aspect of meaning and behavior. It is
impossible to conceive language or life without this affective element. Language
would be a purely deictic form of communication, simply pointing to entities without
the potential for the interaction that is vital to communication – and to the processes
of life.

In the interest of succinctness I will often call the descriptive set alpha factors
and the affective, motivating set the beta factors. Together these constitute a very
robust set of factors that mesh to provide a working structure of word meaning, one
of the foundational elements of language. I will describe the beta factors a little more
briefly in the interest of space. We will see later that these have an enormously
significant importance beyond language.

7.1 The meaning of d.
This is one of the most central of the affective factors with its simple polar senses of
positiveness and negativeness, attraction and aversion, good and bad. It is
also one of the most convincing in terms of its polarity. Some of the key iconic
words are:

Dance, danger*, dare, dark*, darling, daughter, dawn, day, dead*, deaf*, debt*,
decay*, decent, decide, decline*, decrease*, deed, deep, defeat*, definite, deity,
dejected*, delete*, delight, deliver, deny*, depress*, deprive*, descend*,
deserve, desire, despair*, destiny, deteriorate*, develop, devote, diamond, die*,
different, difficult*, dignity, diminish*, dinner, dire*, dirt*, disappear*,
disappoint*, disaster*, discover, disgrace*, disgust*, display, distress*, divine,
do, doom*, doubt*, down*, dread*, dream, dress, drink, drop*, dull*, dumb*,
duty, dwell.

The affectively neutral non-iconic core words are much less numerous: date, deal,
dig, dish, distance, dog, door, drag, draw, drift, drip, drive, dry, due, dust
.

Content. The most outstanding feature of this list is the dominance of affective
content compared with most of the words in the sections we have already examined.
(The r section is a notable exception).

A number of the words that I have identified as affective and iconic may seem
questionable. This applies particularly to dawn, day, deed, deep, dinner, dirt, do,
drink
. Superficially these may appear indistinguishable from the non-iconic words but
I believe they are further examples of an inherent existential affectivity that is virtually
absent from the non-iconic d words. The distinction is fine in some instances, but
sustainable in my view.

In addition to the d words, it is obvious that this factor exists with the same
impact in innumerable words that do not begin with d.

Polarity is clearly very defined. Even if we extract the effect of the Latinate de- and
dis- words the polarity is clear.

Motivation is very difficult to determine as nothing in the articulation, apart from the
presence of voicing, seems to be relevant. The close to frontal position of articulation
runs counter to the principle that affectivity is generally associated with backness of
location. The motivation is an enigma. Frontness also applies to the next three
factors but other motivating features are evident there.

7.2. The meaning of v and z: fullness, emptiness.
One factor operates with an equally strong natural polarity in both of these sections. I
have designated it as fullness and emptiness with the terms generally applying in a
figurative sense.

Word list:
Vacant*, vacuum*, vagina*, vague*, vain*, valley*, value, vanish*, vanquish,
vantage, vapour*, vary*, vase*, vast, vaunt, veer, vehement, veil*, vein*, velocity,
veneer*, venerate, vengence, venture, veracity, verb, vertical, very, vessel*,
vestige*, vex, vibrant, vicarious*, vicissitude*, victim*, victory, view, vigour,
vindicate, violence, virgin, virtue, vision, visit, vista, vital, vivacious, vivid, vocal,
voice, void*, volume, vomit (v.)*, voracious, voyage, vow, vowel, vulgar*,
vulnerable*, vulva*.

There few core words that I judge to be non-iconic: vegetable, village, vine.

Zany*, zap, zeal, zen*, zenith, zephyr*, zero*, zest, zig-zag, zing, zip, zombie*, zone*,
zonked*, zoo, zoom.

Sub-factors:

........1. Salience: value, vantage, vast, vertical, very, view, vision, vista; zenith.
........2. Fullness of positive values: value, venerate, venture, vibrant, victory; zeal.
........3. Physical or figurative fullness: vaunt, velocity, vindicate, vigour; zap, zest, zoo.
........4. *Emptiness, insubstantiality, lack, absence: vacant, vague, vain, valley, void;
............. zany, zen, zero, zombie.
........5. *Physical and figurative emptiness that is also a capacity to contain;
.............containment
: vacuum, vagina, vase, vein, vessel, voracious, vulva; zone.

Content. There is a well-developed and rich factoral scope in these sub-sections. The
last sub-factor is more complex than most but has a high utility-value in the factoral
analysis of word meanings. There is some similarity between this polar aspect of this
factor and two later factors, q and u but all have specific functions in representing
contrary aspects of cognition that range between distinctive poles. There may also
seem to be some overlap with the g factor, abundance/paucity but the words of the
sections indicate a clear distinction.

/V/ is identical with /f/ in articulatory terms except for the presence of voice.
There is an interesting parallel between the two in that the fullness of v parallels the
firm surface of the first f sub-factor and its emptiness parallels the last two.

The z section displays an almost identical polarity to v but lacks the fifth subfactor.

Motivation. Both sounds are voiced and have a distinctive vibrancy that is appropriate
for the semantic factor. Even voidness vibrates with affect. The two poles are as
interdependent as the Chinese yin and yang, darkness and brightness.

7.3. The meaning of s: body and affect.
As the s section is the largest in the English dictionary I will list only the core words.

Sacred, sad, safe, sake, salt, same, save, say, scar, scarce, scare, scream,
secret, see, seem, self, semen, send, sense, separate, serious, sex, shape,
shame, share, sharp, she, shift, shit, shock, short, shout, shut, shy, show,
shock, shout, show, shy, sick, side, sign, silence, simple, sin, since, sing,
sister, sit, size, skill, sleep, slip, slow, small, smell, smile, snatch, snot, sneer,
sniff, so, soft, son, soon, sore, sorrow, soul, sound, space, speak, speed,
spend, sperm, spit, spoil, spread, stain, stand, start, starve, stay, steal, steep,
stench, step, stiff, still, story, straight, strain, stretch, strong, struggle, stupid,
substance, such, suck, suffer, suggest, sure, surface, sweat, swell, sweep,
sweet, swim, sympathy
.

Core non-iconic words: sail, sand, scatter, scene, school, sea, season, seed,
set, seven, sheep, sheet, shell, ship, shoe, shop, shore, six, sky, smoke, snow,
soil, spare, spill, spot, spray, spring, square, star, state, stone, storm, stream,
street, summer, sun

Sub-factors:

........1. Bodily-felt affect: sad, safe, scarce, scare, secret, serious, shame, soon, sorry,
............ stupid, suffer, sure, sympathy.
........2. Bodily and psychological actions and states, personal states: say, scare,
............ scream, see, share, shout, shut, smile, stand, steel, struggle, suggest.
........3. Sensations: salt, see, sharp, shock, silence, smell, soft, sore, stiff, strain.
........4. Associated with the body: semen, sex, shit, sperm, stand, suck, sweat.
........5. Affective neutrality: same. (This sub-factor has little representation in the
............section).

Content. This and the following related section, t, are somewhat problematic in that
they do not fit as neatly as the rest among the beta factors. The description I have
chosen for s is body and affect. The physical body, the sensory body and the
affective (emotional) body are all strongly represented. The first impression on
perusing these words may be of sheer disparateness. At first sight it appears almost
impossible to extract any common semantic thread. There is however a cue in the
first four words that is shared with many other words in the section such as save,
scarce, secret, serious, shame, share. I have called this thread bodily-felt affect
referring to a variety of distinctive and sometimes subtle affects that are essential
features of these words. They are termed bodily-felt because they have bodily
manifestations that are intrinsic to their sense. Few of these are overt emotions (sad
and sorrow are exceptions). The bodily aspect evident here is consistent with the
bodily character of all the factors. It is more overt in the next three sub-factors. The
semantic pattern that emerges seems to prefigure something very contemporary, the
idea that body, affectivity, mind and the senses are an indissoluble unity. There is no
divide between body and brain, body and mind. Affect and sensoriness are powerful
binding forces and the full unity is operative in action and behavior as well as in
language.

Motivation.
/S/ is voiceless sibilant continuant formed in the front part of the vocal apparatus but
less frontal than the labial sounds. No motivational significance can be read into this.

Polarity. We can postulate a virtual pole, affective neutrality that is manifest in all the
words I have specified as non-iconic, which I have done, but I am not confident that it
is a valid pole. Another possibility is that the pole for s lies in the t section that, as we
will see has a close association with s in articulatory and semantic terms. I will review
this possibility at the end of the next section.

7.4. The meaning of t: tactility.
This factor is somewhat more straightforward.

Word-list:
Table, tackle, tail, take, talk, tall, tamper, tangent, tangle, target, task, taste, tattoo,
taunt, taut, teach, tear (v.), tease, tell, tempt, tenable, tenacious, tend, tension,
test, texture, than, that, the, theft, their, then, thence, there, these, thick, thin,
thing, think, thirst, tidy, this, thistle, thorn, those, thrash, threaten, throb, through,
throw, thus, thwart, tickle, tight, tingle, tiny, tip, tired, titivate, title, to, toe,
together, toil, tone, tongue, tool, tooth, top, toss, touch, tough, toy, towards,
trace, track, trail, trait, trample, tread, tree, tremble, trial, trickle, trigger, trouble,
true, trust, turn, tweak, twine, twinge, twist, twitch.

Non-iconic core words: tame, tear (n.), ten, thank, thousand, three, throat, tide,
time, today, too, town, tribe, twenty, two.

Sub-factors:

........1. Tactility: take, task, tear (v.), tease, test, thrash, throw, tickle, touch, twist.
........2. Sensation: taste, texture, thick, thin, throb, tickle, tremble, torture, twinge.
........3. Pointing or deixis: that, the, their, then, there, these, this.
........4. Uprightness: tall, thorn, tip. (The prime exemplar is stand in the previous
....... .section).

Content. This factor is connected to the s factor by their common marked somaticity.
All the factors are intrinsically body-related in one way or another, but the connection
is especially cogent in these instances. The iconicity of most of the t words is
indisputable but several may appear questionable. I judge these to be iconic for the
following reasons:

........Table – a key feature of a table is that its top is at a level where hands can
................... operate with ease.
........Tail...even our tail-less species can empathize the tactility of having one.
........Talk..the touch is in the tip of the tongue, as with tell.
........Teach – this is optimally a hands-on activity with much pointing and
....................more talking.
........Think – a similarity between the manipulation of physical things with hands
....................and the process of thinking is a possible basis for iconicity here.
........To..... the tactility here may lie in the most basic meaning of approaching
....................and reaching (touching).

A unique feature of the section is the number of deictic words. Tactility, or the
not-quite-touching of pointing, is an appropriate analogue for deixis.

Motivation. The motivation for the association of /t/ and tactility is one of the most
transparent and convincing. The light touch of the tip of the tongue on the alveolar
ridge behind the upper front teeth without the weighty presence of voice is a very
appropriate analogue.

Polarity. As with s, there is no obvious polar region in this factor. The main possibility,
raised above, is a polar relationship between the t and s factors. What seems
feasible is that the affective and existential elements of s (as in sacred, sad, safe,
same
) and the tactility of t (take, talk, tangle, task) with its negligible affectivity are in
a polar relationship. Although this does not seem particularly plausible it is the best
available solution unless we conclude that polarity is not mandatory in the factoral
system. This, too, is a credible stance. There is no option but to leave this untidy
loose end.

7.5, 7.6. The meanings of g and h: generosity and possession.
If there is doubt about the polarity of the s and t factors there is none in these
sections. There is a much more significant overlapping here in that each section
contains conspicuous examples of the factor that is dominant in the other. The factor
that emerges in the g section is abundance and generosity and that in h is the
polar possession and self-interest. This is the classic existential dichotomy of
giving and having, open-handedness and selfishness. These are powerful forces
both in our lives and in society. They are motivators of personal action, makers of
history in both its dark and bright episodes, and shapers of religion and spirituality.

Core words.
Gain*, game, gap*, garden, gasp*, gather, gay, general, generous, get*, gift,
girl, give, glad, glisten, gloom*, glory, glow, goal, god, gold, good, grab*, grace,
grand, grasp*, grass, great, greed*, greet, grief*, grip*, ground, group, guess,
guest, guide, gut.

Non-iconic: go, govern, gray.

The presence of the semantic factor is clear. Most of the asterisked words, with the
exceptions of gasp, gloom and grief, have a distinct kinship with the h factor, but
these three have a generic negativity that is also evident in several core words in the
other section. The g/h polarity is obvious.

Sub-factors:

........1. Abundance, availability: garden, gather, generous, gift, give, god.
........2. Generosity, goodwill, care: game, gay, glad, good, great,
........3. Productivity, effectiveness, affordance: grace, guess, guide.
........4. *Paucity, deprivation: gasp, gloom, grief.
........5. *Seeking, grasping: gain, get, grab, grasp, greed, grip.

Habit, hall, hand, handle, happen*, happy*, harm*, harvest*, hate*, have, he,
head, heal*, health*, heap, hear, heart, heaven*, hell*, hello*, help*, her, herd,
here, hero, him, history, hold, holiday, holy*, home, hope*, host, house, how,
hug, huge*, human, humble*, humour*, hunt, hurt*, husband
.

Sub-factors.

........1. Possession, acquisition, territoriality: handle, have, hear, here, hold, house.
........2. Want, desire, appetite: habit, hug, hunt.
........3. Self-interest: hand, happy, heal, health, heart, heaven, holiday, home.
........4. Personhood, subjectivity: he, her, him, human, husband.
........5. *Loss, detriment, hostility, threa): harm, hate, hell, hurt.

Content. We have a similar picture to that in the g section. The factoral meaning is
crystal clear in habit, handle, have, hear, here, hold, home, hunt. The same generic
negativity that appears in the g section is evident in harm, hate, hell and hurt. The
other marked words, heal, health, heaven, help, hello, holy, hope, huge, humour are
more consistent with the g factor than with h but some are ambivalent. Happy, health,
heaven, holiday, hope, hug, human, humour
all contain elements of both generosity
and self-interest. The g and h factors are two sides of a coin. The presence of the
deprivation, loss theme implies a more complex polarity similar to that in the m and p
sections.

Motivation: There is a common explanation for the articulatory motivation of the two
sounds and their associated factors. I mentioned earlier that position of articulation in
relation to the front and back of the oral region is a motivating factor in many
sections. /G/ and /h/ together with /q/ and /u/ are articulated near the back of the oral
apparatus, deep in the mouth, a location that they share with c (velar /k/). The factors
associated with these sounds – abundance, possession, uncertainty, displacement
and community
– are all profoundly existential. There is a deep logic in this
association in contrast with the predominance of world-descriptive factors that are
generally associated with a frontal articulation.

7.7. The meaning of j: personal energy.
I have interpreted the factor involved in this section as having just two sub-factors,
personal energy and humour, happiness. The second part seems to be somewhat
anomalous because it is less abstract than the other factors. On the other hand it is a
vital human attribute and so may warrant its factoral status. No polarity is evident
here.

Word-list.
Jack (v.), jealous, jagged, jam (v.), jar (v.), jaunt, jazz, jerk, jest, jet (v.), jig, jiggle, jilt,
jingle, job, jockey, jocular, jog, joggle, join, joke, jolly, jolt, jostle, jot, journey, joy,
jubilee, judge (v.), juggle, jumble, jump, just (a.), just (adv.), juxtapose.

Non-iconic: jacket, jade, jam (n.), jar (n.), jargon, jaw, jelly, jersey, jewel, journal, jug,
juice, jungle, junior, junk, jury.

Motivation. The iconicity is robust but no polarity is evident. The letter j was
introduced only in the seventeenth century but the sound is presumably much older.
In phonetic terms /j/ is described as a voiced palato-alveolar affricate. It is produced
in the middle area of the mouth by pressing the tongue against the palate and
through to the alveolar ridge and releasing it with the emission of air and voice. The
sound is very distinctive and the articulatory action requires more effort than that of
most other English sounds making it an appropriate gestural analogue of the
semantic factor. It is interesting that both the factor and the form of articulation have
similarities to r. This is another small piece of the iconicity jigsaw.

7.8. The meaning of q: uncertainty.
This too is a small section but its factor has a disproportionately important role in the
full system. It emerges unequivocally in its polar form from the core words and occurs
frequently in the meanings of other words. Uncertainty, indeterminacy is the
meaning that can be identified here, together with the polar definitiveness,
distinctiveness, authority. Both poles have an important place in the factoral
system.

Word-list.
Quaint, quake, quality*, qualm, quantity*, quark, quarrel, quasi-, queer, quench*,
quest, question, quibble, quick*, quiet, quintessence*, quirk, quit*, quite*, quiver,
quixotic, quiz, quorum*, quota*, quote*.

There are no non-iconic words. Articulation and meaning have no obvious motivating
connection. This factor is closely associated with the following one.

7.9. The meaning of u: Displacement.
The description I have given to the head factor is a catalogue of woes –
displacement, dislocation, disorder, negation and the associated wrongness,
uselessness, discomfort and disparagement. The counter-balancing pole is
correctness, satisfaction, comfort, normality, order and utility. These all emerge
from the following word-list.

Ubiquitous, ugly, ultimate*, umbra, un-, unable, uncertain, unconscious, under,
understand*, undo, undress, uneasy, unfair, unhappy, uniform*, union*,
unintelligible, unique*, unison*, unit*, universal*, unless, unlikely, unreal, unrest,
unseen, untenable, unthinkable, untie, unusual, unwell, up, upright, upset,
upside-down, urge, urgent, use (n. and v.)*, useful*, useless, usual*, utensil*,
utility, utmost, utopia, utter (v.), utter* (adj.).

Non-iconic core words: uncle, us.

Content. The prefix un- dominates the section with its power to turn upside-down the
positive meanings of words, but the small remainder is sufficient to establish the
credentials of both poles. Ugly is a prototypical form of displacement. Under and up
are departures from a more normal position. The polar union implies a desirable
conjunction. Unique is an exemplary state. Urge upsets the status quo. Use
overturns stasis. The verb utter transfers information from mind to world. This factor
with its polar meaning also has a prominence in the factoral system that belies the
smallness of the section. The disparagement sub-factor is barely evident in these
words but as we have seen it is common in word-final phonaesthesia as well as
having a more general occurrence as a medial vowelin words with a mild or marked
derogatory sense as in fumble, jumble, tumble and innumerable others.

Motivation. The sounds of u, especially the dominant short schwa sound as in ugly,
are uttered in the deepest vowel position in the larynx. This can be regarded as an
appropriate motivation for the primary meaning.

7.10. The meaning of w: existentiality.
This item of the alphabet is puzzling and unique as might be evident from a perusal
of the impressive list of core words. There are no non-iconic core words.

Core words.
Wait, wake, walk, wall, want, war, warm, wash, waste, watch, water, wave, way,
we, weak, wealth, wean, weapon, wear, weary, weather, weave, web, wed,
weed, week, weep, weigh, welcome, welfare, well, west, wet, what, wheat,
wheel, when, where, whether, while, whisper, white, who, whole, whore, why,
wicked, wide, widow, wield, wife, wild, will, win, wind, window, wine, wing,
wink, winter, wipe, wise, wish, wit, with, woe, woman, womb, wonder, woo,
wood, wool, word, work, worm, worry, worse, worship, worth, wound, wrap,
wreck, write, wrong
.

This is as extraordinary a catalogue of items, actions, qualities and important
grammatical terms, the interrogative words, as is found in any section. Almost all are
in the classical core monosyllabic form and have Germanic origins and many can be
traced back to Indo-European roots. I have no explanation for this disproportionate
number of core words. The most surprising aspect of the list is that no specific
factoral theme emerges, not even a hint of a factor except, perhaps, the demandingof-
an-answer of the interrogative words, but they are just a fraction of the total. This
lack is surprising in the light of the emergence of factors in every other section of the
English dictionary (except the miniscule x section). I believe, nonetheless, that there
is a factoral theme.

The clue to its identity is the fact that these are all core words. The common
factor among the core words in all sections is (or should be, but I have tended to
include too many words in this category in order to illustrate factor meaning) that they
are all part of the indispensible verbal base of human communication. Each core
word should encapsulate a concept that has a unique and fundamental place in
human cognition as expressed in words. I have searched for a term to describe this
quality and the best I have been able to find is existentiality, the quality of belonging
to the bedrock of human experience and cognition. All core words should have this
quality. The w section is especially rich in such words as it contains far more core
words in proportion to its size than any other section. The words in the list thus share
a character that runs through all the true core words we have surveyed in other
sections. Conversely this factor is part of the meaning of all other core words.

Motivation.
Motivation is unclear. The articulation of /w/ is quite complex. It involves a protrusive
pursing of the lips followed by an un-pursing with a simultaneous onset of voicing that
takes the form of the following vowel. The place of articulation, the front of the mouth,
is as noted earlier, the location for many consonants that are associated with
descriptive factors. This, however, is not a descriptive factor but one with a profound
affectivity, the subtle sense of belonging to the existential bedrock. The closest
articulatory counterpart to /w/ is /q/ that also has a factor with a deeply existential
sense that could possibly be regarded as having a polar relationship with the w
factor. The one is characterized by indeterminacy and the other by a sense of
unquestionable affirmation (apart from the question words). It is also noteworthy that
the interrogative words in this section could equally appropriately be placed in the q
section.

This does not explain the motivation of the w factor, but in the context of the
web of connections between articulatory forms and factors it has a semblance of
credibility. I regard this factor as the foundation stone of the factoral system.

7.11, 7.12. The meaning of c and e.
This brings us to the end of the factoral saga. Like the section of the dictionary we
have just dealt with these two are also somewhat problematic both in terms of
identifying factoral content and in the nature of the content. The factors that seem to
emerge in both sections have a different character from all their predecessors. They
can be viewed as having a taxonomic function rather than representing substantive
semantic constituents. Under this interpretation they specify, for each word, whether
what it denotes exists in the social life of communities or in the purely physical world.
In computational terms this may be a significant consideration in determining the
meaning of a word. One of the most basic things we need to know, in the instant of
deciding the meaning of a word is whether the entity or event occurs is a human
context or is part of the physical environment. These two factors seem to perform this
function. I have named them belonging to the human (c) and physical (e)
lifeworlds or domains. The term lifeworld is borrowed from the German
phenomenological philosopher, Edmund Husserl. I have employed it because it
indicates that both factors relate to human experience in especially fundamental
ways. Here is my selection of the iconic and non-iconic core words in each section.

Human lifeworld.

Cake, call, camp, can (v.), care, carry, cart, cat, catch, cause, certain, chair,
chance, change, cheap, cheat, cheek, cheer, cheese, child, choke, choose,
church, clap, clean, clever, climb, close, cloth, coat, come, comic, common,
company, compare, complain, connect, conscious, consider, continue, control,
cook, copy, correct, cost, cough, count, country, cover, crawl, create, creep,
crowd, cruel, crumb, cry, cup, cure, cut
.

Non-iconic: cave, centre, circle, claw, clear, cloud, cold, colour, corner, cow,
crack, crash, crush
.

Physical lifeworld.

Each, early, earth, east, edge, effect, egg, either, element, else, empty, end,
energy, enough, equal, even, event, evening, ever, every, exact, except, exist,
external, extreme, eye
.

Non-iconic: easy, eat, effort, emotion, enable, encourage, enemy, enjoy, enter,
escape, evil, exchange, excite, expect
.

Content. The distinction here between the iconic and non-iconic words is more
important than usual. We can see that the non-iconic words of each section fit fairly
neatly under the factoral meaning of the other, producing a large overlap between
them. In the c section the factor emerges strongly with a clear-cut group of non-iconic
words. It is more difficult to make the necessary distinction in the e section where a
number of words such as each, else, enough, even, ever, exact, except seem to
straddle the boundary. The c factor is robust, but the e factor much less so. I believe
this doesn't jeopardize the interpretation. Although the proportion of iconic to noniconic
words is lower than it is most sections the proportionality is still positive.

There is another piece of support that is not fully reflected in the word-list. This
is the large group of words with the Latin prefix ex- in the e section. Many of these
words are important to the English lexicon but most do not count as core words. The
productivity of the Latin preposition ex- = out of, from within is indicative of the
importance of this basic notion in the lexicon. The Latin co-, con- / com- = with,
together with, in the company of, together
is a polar concept with an even larger
influence. I believe this adds a degree of affirmation of these polar factors that have a
strategic position in the configurations of factors that form the framework of the
meanings of words.

Motivation. Can we discern motivation for these factors? I have already noted that
consonantal sounds formed at the back of the oral cavity – the velar /k/ of the c
words, /g/ and /h/ – all have associations with semantic factors that share a powerful
human, affective character in clear contrast with the majority of the core words that
begin with consonants formed in the frontal region such as m, p, b, f, n, l. that have a
descriptive character. This backness compared with the frontality of the descriptive
consonants seems to be a powerful motivating factor for c. Something similar seems
to prevail with e. Being one of the most frontal vowels makes it appropriate for the
meaning, relating to the physical lifeworld.

7.13 Summary and comment for Sections 6 and 7.
This is the full set of semantic factors as I have been able to identify them. What is it
that we have found lying within the lexicon? I believe it is something quite
extraordinary. It has the appearance of a code operating at the heart of word
meaning and forming linkages with the phonological system. It is a code of primitive,
abstract proto-concepts that arguably have a critical function in the formation of
concept and word meaning. They are abstract in that they possess the simplest
content that such an entity could possess while maintaining an ability to represent
aspects of things and events in the mental medium. They are primitive in that they
are based on primal aspects of human experience. With these characteristics they
are unable to represent the world viridically and in fine detail on their own, but they
can provide essential structure in a way that is efficient in neural terms and that
enables supplementary material to fill in the detail to the degree that is necessary.

As such, I believe the set of factors may prove to be a Rosetta stone for
coming to a better understanding of word meaning and its function in the language
system. The stone has been lying face down where linguists have stumbled over it
innumerable times, but they have never had the curiosity to turn it over. If it is such a
discovery, the puzzle of the nature of word meaning that I alluded to at the beginning
of this paper is on the way to being resolved. It is very surprising that this iconic
semantic code is to be found more or less intact within the English lexicon. While the
specific iconicities described above are specific to English and, to some degree,
some allied and other languages, the factoral code has the hallmarks of having far
wider application as discussed in Section 10. In the following section I will describe
the way the factors are manifested to a substantial degree in a totally unrelated
language.

I believe the evidence that the factors have such a central role is compelling
but a sceptic could take the view that these associations of sound and meaning are
epiphenomenal and of little consequence, a mere pervasive sound symbolism, akin
to phonaesthesia, and nothing more. I believe this does not accord with the facts as I
am presenting them. The factors possess an impressive power to explain features of
the language system in a superior way as will be demonstrated further in the rest of
the paper.

We have now seen the full factoral system operating in the lexicon. I have
often referred to the set of factors as a system. It is a system in the sense that this
small set combines generatively to construct the expanding universe of the meanings
of words of all languages (at the structural level). Systematicity is the capacity of an
entity to produce results consistently and effectively to specific functional
requirements. It implies a law-like regime. It is regarded as a hallmark of genuine
elements in linguistics and many other fields of science. Systems are legion in the
biological and physical worlds. One of the tasks of science is to identify them and
their functions and mechanisms and interrelations with other systems. There are also
an ever-increasing number of artifactual systems that operate all around us. One of
the oldest is the alphabet, a comparatively simple system for representing the sounds
of words. Here we have another alphabet, the alphabet of word meaning. The
ultimate test of its systematicity is its capacity to enable the neuronal brain to
construct ‘models' of events and entities in the world that allow the human organism
to think and communicate. This is a very tall order.

But what justifies us in regarding the alphabet of meaning as a system? So far
we have seen the semantic factors operating singly in word meaning and pairs of
factors ostensibly combining in some phonaesthemes. This is a mere glimpse of the
power of the factoral system. In the following chapters this glimpse will expand into a
panorama with some unexpected implications.

But first we need to take stock in order to determine that the previously
unrecognized set of semantic factors is credible. An iconicity involving a so-called
semantic factor has been identified as being associated with almost every letter of
the English alphabet. The credibility of this discovery depends on the soundness of
my intuitive identification of common elements of meaning across the core words of
sections of the dictionary and the actual content that I have attributed to those
common elements. I have acknowledged that there is a lot of scope for different
judgments on matters of detail but I have held that the patterns of iconicity and the
identification of the factors are robust. I submit that this is generally hard to dispute,
given the evidence. With regard to the descriptions of the individual semantic factors
and their sub-factors there is also considerable scope for proposals of variants but
the general outlines have a remarkable firmness. I believe, in sum, that we have a
viable description of an unanticipated phenomenon within language that has
substantial evidential robustness and that will be demonstrated in the rest of the
paper to bring a flood of new light on the nature of word meaning.


annotated contents




8. The semantic factors in Maori.

The investigations that I have described have been entirely Anglocentric but the
validity of the semantic factors requires that they apply to the words of any language.
My investigation of other languages has been limited to several European languages
that descended from their common Proto-Indo-European source, plus Finnish which
belongs to an unrelated family, and Maori, the language of the natives of New
Zealand and an outlying member of the vast Austronesian language group. In all of
these languages I have found quite extensive evidence of the factors as
demonstrated by articulatory iconicity. My study of Maori is the most extensive (see
Lloyd 2013).

The phonological system of Maori is one of the smallest of any of the world's
languages. It has five vowels a, e, i, o, u that can be in short or long form, and ten
consonants h, k, m, n, ng, p, r, t, w and wh. The last normally has a sound close to
English /f/. The Maori /r/ is rather different from that of English. It is described as a
flap and has some similarity to English /d/. But on the whole these sounds are close
to their counterparts in English. The smallness of this phonological repertoire would
seem likely to have a considerable effect on any factoral iconicity in Maori.

Several big surprises were in store in this study because there was no prior
reason to expect the factors to be represented iconically in the Maori vocabulary. The
first surprise was that factoral evidence emerged quite quickly from an examination of
the Maori lexicon along similar lines to that I had undertaken on its English
counterpart – extracting the most commonly used words in sections of the dictionary
and seeking semantic commonalities among them, that is in words with the same
initial phoneme. Well-attested semantic factors were identified in almost all the Maori
sections. The second surprise was that the factors are associated with the same
sounds, or sounds of similar articulatory form, as in English. This need not
necessarily have been the case because articulatory form is generally capable of
bearing different analogical interpretations and potentially becoming associated with
a different factor from those in English. This result suggests that the iconic
associations found in English may be especially robust and may be common in many
languages with similar phonological systems. The last surprising conclusion from this
study was that while the set of factors identified in Maori are by no means a perfect
match with those found in English, they have been useful in helping to clarify some
puzzles and apparent anomalies in the English factoral system and thus contributing
to a better system. Further languages need to be studied to resolve some
outstanding issues.

...........Some key features of the factoral system in Maori are:
........• The same factors as in English occur in most of the sections: m, p, b (in the p
........... section), f (in the wh section), n, l (in the r section), o, a, i, k, v (in the w
........... section), s and t (in the t section), h and g (in the h section), q (in the u
........... section), u (in the ng section) and c with a more substantive meaning than in
........... English in the k section. Some are more robust that others in terms of their
........... support in the vocabulary but these factors have a clear presence in most of
........... the Maori sections. This is a remarkable result but several factors are notable
...........by their absence: extension (y), intensity of effort (r),
...........positiveness/negativeness (d), personal energy (j), existential value (w)
.
........... These factors exist, by necessity, in Maori word meaning but are without any
........... iconic representation.
........• In the Maori h section the two factors that are found in English g and h,
...........generosity and possession, are present in a polar relationship.
........• There is evidence in the Maori k section that the entirely different factors
...........associated with velar and palatal /k/, community and physical intensity are
........... often distinguished in Maori according to the following vowels. The first k
........... meaning in English was given a taxonomic function rather than a fully
........... substantive function. In Maori this velar /k/ is markedly iconic and is
........... associated with a substantive semantic factor.
........• Maori has no /b/ but the factor associated with b in English, bodily
...........roundedness, is clearly evident in the p section alongside words in which the
...........p factor of English, particularity, occurs in that language.
........• The Maori /r/ is associated with the same factor as occurs in the counterpart
...........English section, l – display, manifestness – despite the somewhat different
........... form of articulation. The /l/ sound occurs in the closely related Polynesian
........... languages, Hawaiian and Samoan.
........• Similarly there is no /s/ in Maori but the Maori t section contains numerous
........... words that are strong exponents of both the s and t factors. A surprising result
........... with regard to the s equivalents in this section was the very large number of
........... bodily emission words in the Maori section, as in the English. Among the t
...........words there is a strong representation of words signifying uprightness that
...........was represented mainly but not robustly in the English s section.

I will use the h section of the Maori dictionary as an example of the investigation I
undertook. As I have noted the polar factors abundance, generosity and possession,
self-interest occur in this one section. The Maori words have significant overlap with
those of English but there are considerable vocabulary differences that reflect a very
different culture. The two groups of words constitute a significant proportion of the
most frequently used (core) words. Here is a selection of words that contain the two
factors, g first.

Haere mai – welcome; haere ra – farewell; hakiri – feast, gift; hanga – make, build,
mend; hangi – earth oven; hapu – section of a tribe, pregnant; harahara – excess,
abundance; hari – joy, pleased; hau – air, breath, essence, vitality; hauora – health,
healthy; here – guide (v., n.), conciliate; hiki – carry in arms, nurse; hine – girl,
daughter; hira – abundant, important; hiranga – excellence, importance, significance;
hoa – friend; hoatu – give (to someone), provide; hoki – also, too; hongi – press nose
in greeting; hou – new, fresh; hua – egg, fruit, abundance; huahua – game,
especially birds preserved in their own fat; huanga – advantage, benefit; huhua –
many; hui – meeting, to meet, gather; hunga – group of people
. These are all
convincing exemplars of the abundance, generosity factor. While the first two
phrases possess the factoral character, the core word haere – come, go, move,
become
is affectively neutral.

And now the possession, self-interest, desire group:

Ha – breath, taste, odour, essence; haha – tasty, luscious; haha – seek; hakere –
stingy; hamu – gather, glean; hao – greedy, acquisitive; hao – catch in a net; harau –
grope for; harawene – envy, jealousy; hau – seek; here – to tie, bind; hi – catch with
a hook and line; hiahia – want, yearn for, be in love; hiakai – hungry; hiawai – thirsty;
hihiri – long for; hinengaro – seat of emotions and thought, desire; hiri – earnestly
desire; hokaka – desire, need; hoko – buy, sell, barter; homai – give (to me); hopu –
catch, seize; horomi – swallow; hotu – long for; hu – desire; huti – fish with a line
.

An important feature of Maori is that it has no direct equivalent of have, the factoral
fulcrum of the English section, as give is of the g section. The concept of possession
is complex in Maori. It is expressed in a number of grammatical constructions in
which the factor is the key, but non-iconic, element. The section is particularly notable
for its desire, seeking theme including the different kinds of fishing, an activity crucial
to Maori survival. Two important core words in this section, he – wrong, fail; and hara
– offence, sin
, can be regarded as polar to both factors in which case they are
comparable to core English h and g words harm, hate, hell, hurt and grief.
The result of this comparison is remarkably clear-cut. The section is an
outstanding example of factoral iconicity.

These results, from a language as distant as it is possible to be from the homeland of
English, are unexpected, to say the least. They comprise further strong corroboration
of the validity of the factors and their universality in the meanings of words in all
languages whether or not they are manifested iconically. The absence of several
factors from Maori raises questions about their validity that can only be resolved by
the study of more languages.

A question remains as to the extent of the iconic representation of the factors
in other languages. A superficial examination of several has suggested that it may
not be particularly significant in some. A much more extensive study of the
vocabularies of other languages is required to settle this question. If it is found that
many languages do not possess much factoral iconicity, this fact would have no
bearing on the presence of the factors in the words of those languages or the
significance of iconicity as the feature that has enabled the semantic factors to be
identified.


annotated contents




9. A new model of word meaning.

9.1 Outline.
I have emphasized that the semantic factors alone cannot supply the full meaning of
words. Their function is structural. A set of 20-odd elements of word meaning cannot
possibly describe the full intelligible content of word meanings. This is a problem that
has been recognized in compositional theories of word meaning where it is known as
‘the problem of completers' as described by Jackendoff (2002). Combinations of
semantic features, however extensive, cannot present a full word meaning.
Something additional is always needed. In this section I will seek to incorporate the
semantic factors into a comprehensive model of word meaning together with two
other elements that give it the ability to represent the full meaning of any word.

9.2 The factoral structure.
I have proposed that the factors construct meaning by combining in configurations
that provide the necessary structure of the meanings of many words, especially
core words, as I will demonstrate in Section 11. By structure I mean a combination of
semantic factors that provides a framework for the non-factoral elements that are
necessary for a fully functional meaning. In the previous two sections we have seen
single semantic factors in operation in an iconic manner but I have emphasized that
the predominant occurrence of the semantic factors in words is not iconic. They
operate freely in forming lexical semantic structure without phonological counterparts
most of the time. The articulatory iconicity of the factors is a mere epiphenomenon in
comparison, but a valuable one because it provides concrete evidence of something
that would otherwise be a lot more difficult to establish. The adequacy of this
structure can vary considerably from quite full simulations of meaning by the factors
to a minimal structure in certain types of words in which the other elements described
in this section fill out the meaning. The meanings of words are far too diverse,
malleable and complex for a semantic alphabet to spell them out in full. Other
components are needed. This requires a new model of word meaning with the
semantic factors at the centre.

There are a number of current theories of word meaning none of which have
achieved consensus (Section 12), but it is widely agreed that the nature of word
meaning is compositional, that it is constructed, at least in part, from combinations
of semantic features of some kind. The model I will propose also involves
compositionality. The question is: what are the other components? A variety of types
of semantic elements such as various kinds of semantic features, images and
prototypes have been proposed. My model has the semantic factors at its centre,
operating as specific configurations for each word, with two other quite different
components arranged in what can be envisaged as a concentric manner. We have
already seen something of the nature and character of the semantic factors and that
will be deepened in the next section.

9.3 Primary and secondary isomorphism.
It is necessary to examine more closely the nature of these configurations and how
they are formed by their constituent factors. The configuration that delineates the
structure of a word's meaning, is an isomorphic (having the same shape or
structure, but lacking in detail) representation of what it denotes. The word ‘land' as in
‘the farmer wouldn't leave his land' has a configurational structure consisting of
materiality (m), spatiality (o), surface (f), particularity (p), display (l), possession (h),
abundance, (g), fullness of value (v), existential value (w)
. This is quite an explicit
framework that conveys much of the sense of the word because it is structurally
isomorphic. The word has the same structure as what it denotes as represented in
direct experience. The piece of land, in the context of the sentence, is physically
characterized by the five alpha factors and dispositionally characterized by the four
beta factors. In other words the configuration achieves its effect through an
isomorphic relationship at a lower level between the individual constituent factors
and the features of the experienced world that they represent. If we were on the spot
and saw the piece of land referred to in the sentence the same configuration of
factors would necessarily underpin that experience. This is a primary isomorphism
(of experience) while that of the lexical configuration is secondary or derivative.

The key to semantic isomorphism is an understanding of the way in which
factors and gestalts are represented in the brain. In 10.2 the origin of isomorphism in
somatotopic or body-related regions of the brain (sensory, motor, kinesthetic etc.) is
described. In essence, the regions of the brain that are responsible for our firsthand
interaction with the world are the basis of semantic isomorphism.

With the beta factors this isomorphism (or rather identity in this case) of affect
is a result of the projection of affects onto entities in the world. They are not any less
significant for that. This is the nature of affects: that they impose a ‘coloration' on
things and events as we can see in the example above. This coloration is very
significant because it is the potential basis for judgement and behavior or action in
relation to the entity. However, the semantic isomorphic configuration falls well short
of a full word meaning. Two other components are required. (Note that in the case of
the structural isomorphism associated with direct perceptual experience the
configuration of factors is supplemented by the visual experience of seeing the piece
of land. This itself as we will see is contingent on the factors in their cognitive guise).

9.4 Illustrations of the model
I will examine another word in some detail to introduce the working of the two
supplementary elements of the new model. First, the verb, hold, taking it in its basic
sense of holding in the hand – he held the rope. How do we know what this word
means and how it means, how its sense forms in the mind/brain? Some of the
following analysis may seem odd at first, but that impression should change with
more familiarity with the working of the model.

‘Hold': belongs primarily to the human life world (c), involves materiality (m), is
a verb (a) with two arguments or valencies, a holder and a thing held
, and involves
durativity (y) (the action persists for a time unlike that of kick or put). These are
‘external' factors, information about the word, more than part of its meaning. We
recognize them from the context the word is used in as much as from the word itself.
On the other hand the vital structural ‘kernel' or ‘nuclear' factors are: bodily action (s),
directed externally (n5), tactility (t), contraction (n), possession (h), definitiveness (q2)
positiveness (d) and existential value (w). (Contraction
relates to the shaping of the
hand around an object in holding it). Anyone can testify to the appropriateness of
these features for understanding the word in a typical context. Note the combination
of alpha and beta (the last four) factors and the isomorphism of the full configuration
with the physical act of holding.

If we think about the word's meaning we can identify all the factors as being
present without undue difficulty, but the full configuration does not constitute an
intelligible meaning as we intuit it in a micro-second. The factors are virtually invisible
to us in the instant of understanding the word as we hear it. What we recognize in an
instant is something of a different character – not a bundle of features, but something
that works as a whole. I propose that this is another distinct key element of word
meaning, what is sometimes referred to as a semantic image or a perceptual symbol.
I call this entity a gestalt (dje'stalt). This is a German term derived from European
psychology of the 1920s and 30s that is still in wide use. Gestalts are distinctive
structures or forms in the mind that can occur in any sensory modality and in or
involving affectivity. They have several distinguishing features. First, they are intuited
mentally as a whole. We can grasp them satisfactorily as an entity. Second, they are
composed of parts but the whole is more than the sum of its parts in our
understanding of the entity. The factors are a new means of understanding the parts
of gestalts structurally. Third, they are ‘felt' in a more or less physical way. Although
mental in nature gestalts have palpable physical qualities. This may seem an odd
attribute for something in the mind. The explanation is that it involves body-related
regions of the brain that generate subtle ‘feelings' that are similar to those
experienced in parts of the body in actual experience. Fourth, as mental formations
gestalts are isomorphic with what they denote and represent. Something in the mind
is similar to something on the world. This is less of a puzzle when we recognize the
primary and secondary isomorphisms described above that can bridge the gap
between word meaning as mental entities and the actual experience of holding or the
perception of something being held.

Semantic gestalts have a number of forms related to bodily experience of
entities in the world apart from the commonest visual type that predominates in
concrete nouns with their high level of imageability. Gestalts occur in several
physiological modalities, an attribute they share with their constituent factors (see
10.2). Their capacity to operate in these modalities is especially important in word
meaning. This is an enormous advantage compared with the mainly visual nature of
most traditional kinds of semantic features. In ‘hold' the gestalt is mainly kinesthetic –
it has something of the nature of the feeling of holding (that, incidentally, varies
according to the physical characteristics of what is held, arope, a teacup or a stone or
a person for instance). This is a consequence of the word meaning being linked to
the region of the brain that is activated by the sensations of holding something. The
main links are the t and n factors. But the word also has a distinct
interoceptive/affective aspect – there is generally something more positive about
holding than letting go. Some theories of word meaning propose that perceptual
images
(11.8) constitute the core of word meaning (Barsalou 1999). This is a
consequence of their proponents' primary focus on concrete nouns that severely
limits the application of such approaches to abstract and semi-abstract words.

But what distinguishes a gestalt from its configuration of factors? A gestalt
must have the quality of instant intelligibility as it is a substitute for a full direct
experience. Factor configurations do not normally meet this requirement on their own.
What appears to be necessary is the supplementation of the factors by sufficient
content of direct experience to provide clear intelligibility. An adequate supplement
may be furnished by the sensory enhancement of one factor by the extraction of
sufficient material from the relevant direct experience as stored in long-term memory.
In the case of ‘hold' it is likely to be something of the sensation of holding associated
with the n factor that is relevant to the context, perhaps in the form of a hand-shape.
In the earlier example, ‘land', this constituent would seem to be visual as would
certainly be the case for ‘lake' while a visual or auditory image would apply to ‘laugh'.
In all instances the other factors in the configurations remain active in the gestalt. The
nature of this supplementary component is hard to determine in some instances.

9.5 Encyclopedic meaning.
The third part of the meaning of the word has an entirely different character. This is
what is known as encyclopedic meaning (EM) or world knowledge. There are
numerous possible constituents to the EM component. Factual information that is
held in long-term memory about what the word denotes is a central component as is
the web of inferences that any concept is caught in. Contexts in which a word is
frequently used, other words with which it commonly co-occurs and the ways in which
it can contribute to the meaning of a sentence are also crucial. Synonyms that would
be more appropriate in some circumstances, and various metaphorical extensions
the word may have are also part of EM. What is required is whatever will satisfy the
complexity of understanding words in the many contexts in ahich they occur. In the
present case EM probably comprises basic ‘folk knowledge' about the act of holding
(what can and can't be held, that holding leads eventually to letting go and so on). An
important feature of EM is that, while the gestalt and the factors are pre-linguistic, this
component exists in a linguistic form and can be diverse and extensive as the brain
taps into its associative ‘innernet' for the relevant material. EM is dependent on
individual experience and knowledge and accordingly can vary extensively from
person to person while generally sharing a common core for widely used words. The
contribution of EM also varies greatly across types of words (see below). It often has
a modest role in core words such as hold that have robust gestalts and factoral
configurations but is generally the dominant element in less commonly used
specialist words.

This then is the 3-part model of word meaning that I will use in the rest of the
paper. The parts operate with highly variable proportions across different types of
words as I will describe below.

9.6. Types of word meaning.
Before putting the model and the alphabet of word meaning to work more extensively
in analyzing the meanings of words in Section 10 I need to outline how its operation
varies across types of word meaning. It is possible to distinguish four main types of
word meaning in which the 3-part structure displays marked and well-motivated
variations.

1. Basic-level core words. This type of core words, that excludes those in the next
two groups, is exemplified by habit, happy, harm, harsh, hate, have, heal, hear, help,
hold, holy, home
. These are distinguished by their relatively weak descriptive alpha
factors and the relative prominence of affective beta factors. Their factoral structures
often provide quite explicit sketches of words' meanings, but without making them
fully comprehensible. The gestalts associated with these words have a prominent
affective tone. EM is of modest size, relatively speaking. A large proportion of these
core words, including all closed-class (grammatical) words (here, how), with their
essential functions in the construction of sentences belong to this important category.

2. Image rich words. The second type includes a large group of core words that have,
as a general rule, substantial perceptual content, dominant often perceptual gestalts,
meager affective meaning and robust EMs. Examples are: hail (n.), hair, hammer,
hand, handle (n.), hang, hat, head, horse
. The group includes many important,
familiar items of the visible personal, domestic and natural world with which we have
close physical experience and interaction, such as parts of the body, household
articles, and a vast range of other everyday artifactual objects, familiar items of
nature and so on. The words that denote some of these things can have significant
alpha factoral structures but affective (beta) content is generally sparse. For the
majority, visual gestalts provide most of their semantic content and factoral content is
characteristically quite limited. The non-factoral gestaltic content of such words can
be described as image-like and sensory. EM is an important supplement for most of
the words.

One may ask why these core words lack more effective factoral structure. The
explanation seems to be that they represent a different type of entity, mainly nouns
that denote objects with a complex physical form that can only be neurally
represented by perceptual gestalts.

I have noted that there is no clear boundary between core and near-core
words. Many words that may be regarded as near-core also have the characteristics
as words in these two groups. These are words like hamper, harsh, haul, hoax,
huddle
.

3. Qualic words. This type consists of core and non-core words that denote visual
and other sensory qualities, the so-called qualic words such as green, pink, smooth,
hard, cold, hot, heavy, high
have very limited factoral structures but simple, strong
gestalts with substantial sensory content. EM is normally less significant than in the
first two groups.

The distinction between the three groups that I have described may not be easy to
appreciate without the kind of meaning analysis I will conduct shortly but here are
three examples with their associated kernel semantic factors to illustrate:

...........Harm (n.) (Type 1) bodily/psychological state (s); disorder (u); deprivation (g4);
....................................... detriment (h5); negativity (d2); somatic affect (s)
.
........... Hammer (Type 2) materiality (m); particulateness (p); artifactuality, (m6);
..................................... manifestness (l); tactility (t); utility (u4)
.
...........Hot (Type 3)....... sensation (s3); physical intensity (k).

Harm has a robust factoral description that provides a substantial semantic structure
through its mutually reinforcing beta factors that constitute a large part of its affective
gestalt. Those associated with hammer and hot could apply to numerous words in
their categories but their gestalts are robust, distinctive and decisive.

In spite of these differences, the core words as a whole are characterized by a
close liaison between gestalts and configurations of semantic factors that is absent
from the next large class of words.

4. Specialist non-core words. The so-called specialist words are less frequently
occurring than core words. They cover a vast range and at one extreme their use is
mainly confined to specific groups such as professions, trades and cultural circles.
Some commoner examples are: hallucinate, hamstring, handicap, handkerchief,
haphazard, harangue, headquarters, heckle, hereditary, heretic, hierarchy
. The
dominant characteristics of this type are (i) the meanings are generally more complex
than those of core words. (ii) The semantic factors normally have a minor role that
fails to provide an effective or informative structure. (iii) The semantic isomorphism
that is characteristic of core and many near-core words is absent from most specialist
words. The meanings are predominantly composed of EM although gestalts are also
sometimes evident. (iv) The meanings are predominantly linguistic in contrast to the
other types. Their EM is stored in long-term memory in linguistic form. An important
consequence of this is that the meanings of specialist words are ultimately expressed
in core words with their factoral and gestaltic composition forming much of the
specialist words' meanings. These words thus have a hierarchical dependence on
core words and, on their semantic factors. Their meanings may also have,
nonetheless, a gestalt-like form when they are familiar. This seems to consist of an
amalgam of fragments of key items of EM combined perhaps with semantic features
‘borrowed' from key words from the EM descriptions involved in the meaning.


annotated contents




10. The nature of the semantic factors.

Before investigating how the 3-part model operates in more detail in the next section I
want to explore the idea that the factors that we have seen to have a central function
in the generation of word meaning also have a much wider and more primal
existence. But first it will be useful to elaborate on the basic characteristics that
underpin the operation of the factors. This should eliminate any lingering concerns
arising from the narrow linguistic basis of my investigations.

10.1 Key characteristics of the semantic factors.
The material in the previous sections provides, I submit, a robust empirical case for
the actuality of the factors based on their functions in constructing the meanings of
words in English. What is becoming evident is that they arguably possess a number
of key qualities. All of these are more or less implicit in the ways in which we have
seen them operating. In brief the semantic factors can be described as:

...........• abstract
...........• primitive
...........• comprehensive
...........• finite in number
...........• non-linguistic
...........• innate
...........• necessary and sufficient (at the structural level)
...........• universal
...........• normative
...........• systematic
...........• of modal generality
...........• somatically expressed
...........• grounded

The factors are clearly abstract and primitive to a much greater degree than
the phonaesthemic meanings and other proposed types of semantic features. They
seem to be as abstract as it is possible to get as constituents of word meaning and
they are amenable to natural extensions along lines of entailment or analogy, to the
sub-factors. These qualities provide the basis of their ability to combine in limitless
configurations to form lexical semantic structure.

They are comprehensive in that they are capable of delineating the structure of
word meanings across the whole range of the lexicon. This is a very surprising
quality. No other set of specific features has approached this capability with
comparable economy.

As an alphabet that effects the composition of word meanings at a structural
level the factors need to be of a finite number for their efficientl operation. While
there may not seem to be a theoretical limit on the number of units in a semantic
alphabet, a moderate number makes for efficiency. The factors form a surprisingly
small set.

The factors are non-linguistic. We express their meanings in words but they
are fundamentally different in nature from word meanings or concepts in that they are
its actual substance, elements of dierect human experience, the neural metalanguage
of semantics.

The factors are ostensibly innate. They are not learned in the ways in which
words are learned. They are either themselves built into the structure of the human
brain or there is a predisposition for them to emerge from experience in very young
children.

They are necessary for language in the sense that the system of word
meanings simply could not operate in their absence, allowing for some overlap and
redundancy. The meanings of core words collapse or are changed if key factors are
omitted. This absolute necessity goes much further, as we will see in just a few
paragraphs. It derives from the factors' dimensional character. While necessary, they
are not sufficient for the construction of full word meanings. However they are,
arguably, sufficient for the generation of their structures. This sufficiency is another
very surprising quality.

From their innateness and necessity it follows that they are universal. No
human language can operate without them because they are fundamental to the
linguistic delineation of human experience and cognition.

The factors have a normative function. In their original behavioral domain (see
below) the normativity of the factors is more obvious. They are parameters of the
judgements that motivate response. This is not the normativity of the ‘ought' kind. It is
ontological (alpha factors) and affective (beta factors). In word meaning the
normativity is derived from that of the behavioral domain. The factors warrant the
fitness of words for their uses and context. The contents of word and concept
meaning are the results of judgements of the validity of their application to what they
denote, whether external entities, mental entities or closed-class lexical concepts. As
primal natural parameters the semantic factors provide the basis for judgements of
applicability and validity.

The factors lack uniform sensory modality. Semantic features are sometimes
regarded as sensorily-coded with an emphasis on perception (images). All the alpha
factors have a modal generality. They can apply to a variety of sensory modalities
with touch being especially prominent. The beta factors, of course, have their
distinctive non-sensory modality, affectiveness.

Nonetheless, the factors all have a somatic, embodied character. They are
‘felt' in some physical way in somatotopic regions of the brain, a quality that they
share with gestalts and the phonaesthemes (see 10.2). They are somatic
representations of primitive elements of human experience. If readers scan the full
set of factors in Annex 1 I believe they will acknowledge that each one can be sensed
in a distinctive almost palpable, albeit subtle, ‘bodily' manner.

It follows that the semantic factors are grounded. The insistence on the
groundedness of language has been a prominent theme in the semantics literature. It
stems from the belief that meaning is contingent on a connection of some kind with
what words denote. I argue that the factors are grounded by their specific derivations
from human experience and location in regions of the brain that give them the
capacity to ground meaning by giving words an in-built capacity to pick out what they
denote in the world.

It is not difficult to see that the set of factors is systematic in that it operates
collectively in generating the structure of word meanings in an infinite variety of
combinations. We observe this systematicity, in particular, in the meshing of the two
types of factors, alpha and beta, in word meaning. This underlies their alphabetical,
combinatory capacity.

The supposed comprehensiveness of the semantic factors raises a question as
to whether the factors as identified in English are the full and definitive set. The way
in which we have already seen them operate indicates that they are at least a very
good approximation of such a set. There might seem to be possible gaps in the
existing structure such as up/down, strong/weak, proximal/distant, moving/not
moving, transitory/long-lasting, familiar/strange
but the current set seems to have a
greater primitiveness than some of these. In addition single factors are capable of
providing metaphorical extensions to account for at least some of these apparent
gaps. Further investigation is needed to confirm the exact parameters of individual
factors and the whole semantic alphabetical system. This evident comprehensiveness
does not preclude some culturally determined quasi-factors operating in some languages.

The existence of a set of entities with these characteristics raises the question
as to precisely how they function in language.

10.2 How the body creates meaning.
A critical aspect of understanding the function of the factors is their basis in bodily
human experience. I propose that the factors are associated with specific
somatotopic or body-related regions of the brain in a systematic way. I believe this
is evident from our intimate experience of the meanings of many words in which the
presence of factors is clear, as I have described them. In some core words this bodily
effect is clearly discernible while in others it is latent but distinguishable in semantic
analysis of meanings conducted with knowledge of the factors.

In broad terms each of the factors has an association with at least one of the
following regions of the brain that are well known to neurophysiologists:

........1. Sensory, mainly perceptual and haptic (the sense of touch)
........2. Motoric, representing action and movement
........3. Kinaesthetic, relating to the position and movement of muscles and joints
........4. Proprioceptive, sensations of bodily position and stance or state
........5. Interoceptive, gut feelings, affectivities, emotions, abstract effects.

These areas of the brain are, of course, intended primarily for the control of
behavioural aspects of the body but I propose that they have had a secondary
function in conceptualisation, thought and language from near or after the time of the
emergence of our species. This is supported by the operation of the factors across
both domains. The presence of the somatotopic connections in individual word
meanings is often quite subtle for two reasons: because they are always part of a
larger configuration of factors with different somatotopic connections; and because
the connections of individual factors becomes somewhat attenuated in sub-factors.
For example the haptic connection of the f factor in feel is robust in the physical
sense of the word, but in fix or friend it is much less perceptible, let alone in false.

The alpha factors relate predominantly to the first three regions and most of
the beta factors mainly to the last. Regions 3, 4 and 5 are less widely recognised
types of somatotopic connection in word meaning. Kinaesthesia is predominant in the
n, y and k factors (near, nest, yawn, yard, kick, kill) while proprioception is prominent
in some words in many verbs in which there is bodily involvement (dance, follow, go,
jump, pray, run, sit, stand, take, throw
). Interoception is an essential constituent of
many words, sometimes prominent and sometimes quite subtly present. It is
prominent, for example, in danger, desire, value, victim, game, grief, heal, harm,
joke, queer, quite, un-, use
. These are clearly affective words but an important effect
of interoception is that it has a key role in the vast number of more or less abstract
words in which it is less prominent, but it provides an integral element of meaning
(appear, arrange, avoid, behave, believe, cause, confuse, decorate, do, expect, help,
know, learn
).

My inclusion of abstract effects under interoception may seem debatable
because fully abstract words have little obvious affective content. My justification for
its inclusion is that in core words of this kind such as experience, freedom, idea,
imagine, interest, listen, meaning, mind, normal, opinion, property, real, secret, state
(n.), think, theory, true, will (n.), wisdom
– words in which there is negligible physical
content and little overt affectivity – there is nonetheless a residual, quite hard-todefine
affective constituent. It can be described as a sense of existential significance
that is perhaps sometimes a combination of the v2 factor, fullness of intrinsic value,
s1, somatic affect
and w, existential value. This aspect of word meaning is
indispensible to language and I propose that it can justifiably be located under
interoception.

One of the foremost conclusions that I have drawn from my investigations of
language is that the somatotopic associations of the semantic factors represent a
previously unrecognised property of language that is central to its primary functions.
This is the ability of the factors to operate as a bridge between the neurons of the
brain and the meanings of words as we experience them. In other words the factors
are instrumental in the construction of meaning in the human mind. If this hypothesis
is valid it provides for the first time a coherent explanation of the generation of
meaning utilizing the dual strands of the alpha and beta factors.

These conclusions should be able to be confirmed by neurophysiological
experimentation. (See Section 14). However, this would seem to require considerable
technological and interpretive advance on what is currenty available because the
somatotropic effects are complex and subtle in many words.

The direct somatotopic connection of the factors with the brain eliminates the
notorious homunculus that has been hard for other accounts of word meaning to
avoid. The body of the language user itself builds the essential structures of the
meanings of words. Much hangs on the hypothesis that I have been outlining. I
believe the following sub-section provides it with strong support.

10.3 Before semantics and human cognition.
I have indicated that the language functions of the semantic factors are a
consequence of their more fundamental functions in human perception, cognition and
behavior. In this section I want to explore the possibility that we can view them as
operating within a much larger scenario as necessary constituents of cognition. My
proposal is that they are also fundamental to the world of animal cognition and
behavior and, in fact, in the interaction of any organism with its environment.

In the human domain the factors also have, in addition to their role in
language, a basic function in motivating human behavior that I have alluded to. They
are not indigenous to language; they were appropriated by language as it emerged in
the prehistoric era. In human behaviour the alpha factors operate as necessary
structural elements of experiential gestalts in identifying entities and the beta factors
function as motivators of appropriate responsive action (doing nothing different most
of the time). I propose that the factors are fundamentally biological and
fundamental to biology
. How can this be?

I will explore the possibility that they have a parallel function across all
species. The ultimate scope of this proposal is that the activity of any organism from
the dawn of evolution is governed by the biological factors that are almost identical
to the set that operates in the human domains of behaviour and language. They are
essential to the physical operation of the biological systems of organisms. Again, how
can this be?

Perceptions or sensations of any kind have the function, across all species, of
identifying entities and situations that are of interest to a perceiver, and of motivating
the responses that are favorable to the organism's well-being. Sensation and
response
(S&R) is the term given to this fundamental disposition of organisms. For
almost all organisms S&R is an autonomous operation that requires no deliberation
or decision. The process is genetically determined. From the beginning (or close to it)
even the most primal single cell organisms had some means of registering aspects of
their surroundings in order to respond in appropriate genetically determined ways.
But what underlay this process? My contention is that the two kinds of factors have
always operated in supporting appropriate adaptive responsive behaviour.

Let us take the classic example from much later in evolution of a toad sitting
and waiting for a cockroach to come within reach of its lightening-fast tongue.
(Simmons and Young 2010). The toad must register the spatial position and the
appropriateness of the object by its size, form and movement in a way that motivates
its innate tongue-flicking response. The mechanism is genetically determined, but
there is, I propose, a set of natural dimensions within which genetic mechanisms
must operate. These are the biological factors. The cockroach has physical features
such as materiality, particularity, size and possibly other topological features and a
gestaltic character of form and movement that are all associated with its recognition
by the toad. It must have a mechanism to trigger its cockroach-catching behavior.
This S&R mechanism must also have parameters to determine levels of certainty and
uncertainty
as to the appropriateness of the object and the possibility of capture.
Others such as positive/negative, availability, self-interest, possession (capture), may
be in operation here although obviously far beyond the ken of the toad. These
parameters can be viewed as dimensions within which the innate mechanism can
operate and the toad can feed, live and reproduce. What the toad responds to
directly is a gestalt, but the sequence of events could not exist without its factoral,
dimensional framework. Dimensions are inescapable. It is not difficult to envisage
how other organisms also must operate within precisely the same type of
dimensional frame, the frame of the factors
.

(This description of the factors in operation may seem too obvious to be
interesting. The factors are very mundane entities. This is however their very point.
These everyday parameters of human and ranid experience operating in the brain,
denoting entities in the world and stamping them with appropriate affectivities, are a
crucial aspect of organismal behavior).

The cockroach has a defense mechanism that makes it hard to capture, hairlike
structures that are extremely sensitive to air movement such as that caused by a
toad's speeding tongue. When they sense the direction of such air movement they
scuttle along the appropriate escape route. A puff of air with the appropriate physical
characteristics triggers its escape action. This response too is determined by
inheritance. The relevant neurons in the insectʼs brain operate autonomously within
an appropriate configuration of factoral dimensions, something like physical intensity,
definitiveness, displacement, negativity
. The outcome of this eternal battle for survival
and nutrition depends not only on the factoral/gestaltic mechanism but also on
precise physical parameters of the encounter and the fitness of participants.

A toad has quite a complex brain that enables it to make its judgments, and a
cockroach a much simpler one, but it is arguable that neither brains nor neural
systems are necessary for the use of appropriate factoral interactional frames. A
primitive mono-cellular organism such as an amoeba needs to sustain itself by
locating nutriment in the form of ingestible pieces of material and consuming them.
This requires motility, which an amoeba has, and the means to distinguish, within its
environment, zones in which it can thrive from those that are poor in nutrients. It must
then distinguish edible from inedible material with is very basic ‘sensory' equipment,
and it must have the motivation to capture and ingest suitable particles. The
organism must interact with its environment in a manner that is consistent with its
survival. Again, this interaction has a genetic prescription, but it would be inoperable
without the dimensional framework. No factoral framework = inertness and
inanimacy. The framework must possess the parameters that enable the organism to
distinguish what is in the interest of its survival from percepts' physical qualities and it
must provide the motivation to implement specific behavior patterns that are required
for its operation as an organism. The alpha and beta factors seem to offer something
very close to the framework that is necessary.

This is a miniscule window on the vast panorama of organismal existence. What
these examples illustrate is that the factors work at a universal, biologically primal
level. They can be regarded as dimensions of the space of the interaction of
organisms with their environments
. The factoral dimensions underlie S&R
mechanisms as conditions for their operation at the interface of organism and
environment. These conditions are rigidly law-like. They are part of the definition of
the possibility of life. While genomes largely determine how organisms behave, the
factors arguably have an equal primality in that genomes depend on them in order to
activate patterns of behaviour. Genes and the framework are locked together
nomically, by natural law, in the operations of life. Whatever the genomic structure of
an organism, a genetically-based function in interaction with the environment requires
the biological factors with the affective factors – as dimensions – having a vital
motivating function, providing meaning, in effect. Primal life processes and the
factoral framework came together at the beginning of life on earth because the
factors are prerequisites for life. The framework would not exist if the world were just
rock and a flux of chemical materials. Its ontology is interwoven with that of genes.

In all organisms and creatures until very recently in evolutionary time the
factoral dimensions have operated purely autonomically in cellular and neural
interactional mechanisms. An important feature of human beings and probably some
of our precursors is that the beta factors have come out of the dark, as it were, and
entered conscious awareness. This has enabled our species to assume a degree of
control and influence over these powerful psychological forces in a way that has
added enormous scope to human interaction with the environment. This in turn, with
the advent of linguistic communication, has enabled us to harness these forces in
determining our collective destinies in ways that have shaped human history.

I have described the biological factors as dimensional in nature but can we determine
their ontology more precisely? We need to consider the two types, alpha and beta,
separately. The alpha factors are simply the obverse of biologically significant
topological parameters of the physical world. They are those parameters as
experienced through the sensory organs of any organism. But the beta factors do not
have any counterpart in the physical world. They are parameters of the selfsupporting
reactivity of organic systems and as such form part of the definition of life.
Any life structure that must react to its immediate environment in order to preserve its
integrity and perpetuate itself by reproduction must operate, it seems, in the
framework of the alpha and beta factors. This would seem to be the case in any
possible natural world. This gives a cosmic dimension to the necessity that is
characteristic of the factors.

There is an important aspect of these vital interactional mechanisms that I have
glossed over. This is the precise manner in which organismal systems employ the
factoral system. How do they correlate a range of dimensional factors that apply in a
particular situation such as a cockroach sensing a puff of air with specific
characteristics in advance of a toad's tongue? The cockroach is responding to a very
specific sensory gestalt sensed on a particular part of its body in a particular situation
– being in the open where it is exposed to risk. The insect has an inbuilt
predisposition to react to this gestalt in a particular way. (These are very particular
insects!) The gestalt, like all gestalts, has parts that include specific dimensional
factors. The question is, how does the insect correlate or compute the gestaltic form
from the factors and any non-factoral components of the sensed gestalt? The only
feasible answer is that the insect's neural system, its tiny brain, contains a
computational algorithm that is genetically designed for this purpose. Leaping
forward to human brains, similar neural mechanisms, but on another scale of
multiplicity, complexity, variety and scope, are responsible for the computation of our
responses, every microsecond of our conscious lives, to what our sensory and
cognitive attention detects whether it be simple sensations, elements of language,
memories, concepts or spontaneous affectivities. Hearing or seeing words triggers
very specific algorithms that generate meaning in ways that I will examine in the next
section.

There is value, I suggest, in regarding the meaning-generating operations of the mind
as working on two connected levels, perceptual/cognitive and conceptual/semantic.
At the perceptual/cognitive level the infinite detail of the world is directly
available to the brain to the degree required by any species in order to be
able to interact with objects and creatures in which they have an interest. It is
profoundly embodied and significantly affectual. Embodiment establishes the
physical connections of brain (and mind) and world. Affect determines organisms'
stances towards entities that motivate the interaction of individuals with their
environments. In most pre-human species affect is purely autonomic, a geneticallybased
evolution-determined evaluative and motivational mechanism. This requires
meaning in the sense of knowing how to respond to whatever is encountered. This
modality is information-rich but the factors have a critical function in the dimensional
structure. In our species perception is sub-served by conception and cognition. It has
access to a reservoir of information that influences response.

The conceptual/semantic level is peculiar to our species. Conceptualization and
semantics involve the ability to form and use concepts free from the physical
presence
of what they denote. These faculties are resourced by perceptual and direct
experience of the world but the brain abstracts from these resources the material it
can use in neurally tractable ways in combinatorial thought and language.
Embodiment and affect retain their full power. Affect in thought and language is no
longer autonomic. It operates sub-consciously most of the time but is available to
introspection. In language and thought it is the essential engine that motivates
meaning, but there it operates in a ‘virtual' mode, retaining its force but abstracted
from direct effect in the physical world.

This excursion into biology reveals the apparent extraordinary prehistory of these
constituents of our every word and it provides a useful background for investigating
how the factors operate in generating the meanings of words.

10.4 Some philosophical background to contemporary linguistics.
It is timely to discuss, albeit very schematically, how the factors accord with some
important streams of philosophical thought, specifically Rationalism and Empiricism,
that are very much alive in contemporary linguistic thinking. Both have had a
profound influence on linguistic thought over the last century. These are polar
philosophical approaches to the understanding of human cognition and language.
Both have their origins in the great seventeenth and eighteenth century philosophical
disputes involving Hume and Locke as key representatives of Empiricism and
Descartes and some later, mainly French grammarians and philosophers, as
adherents of Rationalism. The central difference in these approaches is that
Empiricism stresses the role of brute sensory experience (including mental images),
and, in the case of Hume, deep-seated characteristics of human nature, in the
functioning of human cognition, the understanding of the world and our systems of
belief, while the Rationalism of Descartes holds in prime place the capacity of
inherent rational principles to provide objective knowledge of the world through the
application of innate logical principles.

The contrast between the thought of Descartes and Hume is particularly stark.
That of Descartes is marked by the priority he gives to intellect, reason and discipline
and clarity of thought over the role of sensation and experience. Hume's philosophy
places a much higher value on the contribution of experience to human knowledge –
‘philosophy cannot go beyond experience' – and he sought to rationalize evidently
profound contradictions of human nature. In the process he investigated the
operation of the human mind in considerable depth.

These are polar approaches but it is obvious from a contemporary viewpoint
that some form of correlation of the two approaches is likely to provide the optimum
solutions. Nonetheless the distinction between them remains potent in contemporary
linguistics.

In the recent history of linguistics Rationalist thought has had the most notable impact
through its advocacy by Noam Chomsky and through his Generative Grammar that
has assumed a commanding position in linguistics. Chomsky's Cartesian approach
was sourced mainly through a school of post-Cartesian French grammarians and
other thinkers deeply influenced by Descartes. Some key Rationalist themes in
Chomskyan thinking are: a de-emphasis of behavioural, performative aspects of
language; an ‘internalist' focus on the mental nature of language as self-contained;
the dominance in sentence meaning of rule-governed structure (as opposed to the
contributions of semantics); the creativity and productivity that language is held to
derive from the same source; the innateness and universality of the structural
principle of language, Logical Form; and the consequent ability of young children to
learn to use language in ways that utilize innate knowledge. These themes culminate
in the notion of a Universal Grammar that is the essential organizing principle of all
languages. Chomsky's generative grammar has had a spectacular success through a
series of radical revisions over the last forty years but it continues to have many
critics and remains controversial.

Empiricism on the other hand has had a very chequered history over this period.
It dominated linguistic thought for several decades of the first half of the twentieth
century in the form of what is known as structural linguistics, a largely classificatory
and descriptive enterprise that fed on increasing awareness of the vast global body
of ‘native' languages that were new to linguistics. In parallel with this activity
Empiricism was the guiding spirit in Behaviorist Psychology with its emphasis on the
experimental study of organisms' responses to stimuli and its rejection of reliance on
non-observable, mental sources of behavior; and also in the philosophical school of
Logical Positivism or Logical Empiricism with its central belief that truth and meaning
must be derived ultimately from the sensations of human experience. Both of these
movements waned under the effects of Rationalist-inspired attacks with Chomsky in
the forefront in the case of Behaviorism.

Over recent decades the influence of Empiricism in linguistics has been more
dispersed and less well-developed than Rationalist-inspired thinking. It has no single
commanding figure. The important Cognitive Linguistics approach to language has
adopted a number of principles consistent with Empiricism such as the notion of
embodiment, that our bodies shape our experience including the experience that is
deeply manifested in semantics. In contrast to Chomsky it adopts an ‘externalist' view
that highlights language's relations with the world. An area more central to this paper
is the issue regarding whether sensory modalities are central to the mental
representation of word meaning or whether it operates predominantly in an amodal,
fully abstract coding system. The psychologist, Laurence Barsalou with his
Perceptual Symbol System and philosopher Jesse Prinz have been prominent
advocates of a fully modal form of cognition (see 12.8) while Zenon Pylyshyn (1984)
has vigorously promoted the amodal approach. This is clearly a fundamental issue in
understanding how word meaning operates but little resolution appears to be in sight.
Both these approaches have fairly obvious and serious drawbacks. Barsalou's
System operates quite effectively with words with concrete meanings but can only
accommodate abstract words through some complicated and barely credible
explanations. The amodal theory fails to recognize the now widely acknowledged
existence of mental imagery. In Section 12.10, after reviewing a number of positions
on the nature of word meaning, I will outline the ways in which the semantic factors
and the 3-part model appear to be able to address these issues, among others, and
to form some reconciliation between the two approaches in linguistics. In the next
section I will provide a demonstration of the effectiveness of the semantic factors as
an alphabet of word meaning at the structural level.

To conclude, Logical Form is the underlying principle of syntax and of the
generation of meaning in language for Chomsky. It may be clear by now, and will be
explained at greater length in Section 15, that this position is very different from that
taken in this paper. This approach is intrinsically Empirical. Word meaning occupies
the central position in the operations of language. Syntax and Logical Form are
derived in large part from the experience-based meanings of words with their
inherent combinatorial potential. It is meaning rather than syntax that is intrinsically
biological and innate.


annotated contents




11. How to spell the meanings of words.

11.1 Formulas of word meaning.

One of the innovations introduced in this paper is a succinct formula-like presentation
of word meaning with the semantic factors forming the structural core and gestalts
and encyclopedic meaning as adjuncts that need to be added by the reader of the
words. In the formulas I use the letters that we have found to be associated iconically
with specific factors in English to represent them. No iconicity is implied, although it
often exists. These are more like chemical formulas than mathematical ones. They
specify the main constituents but not the composite product with its intrinsic
properties. That is the function of gestalts and EMs.

The ‘letters' of meaning configurations are not abstract arbitrary symbols like the
letters of the alphabet. They are elements of human experience – abstract but
packed with denotational scope. This gives them the power to form the structural
nucleus of the meanings of words. In the formulas the relevant gestalt and EM is
attached at each end like a head and a tail but they are integral to the full meaning.
Thus the meaning of mean (v.) as in What does this word mean? can be presented
as:

*[c, m7, a2, y, (s2, n5, (p:p), i4, g3, p, l, v2, q2, s, w)]E

in which the asterisk represents the non-linguistic gestalt that can only be described
verbally in a general way, and the E stands for the relevant encyclopedic knowledge
that each person brings to the understanding of the word. Each of the letter symbols
is a factor or sub-factor. The formula looks cryptic but it will be easy to interpret by
the end of this section.

Cautionary note: The analyses below are part of a work-in-progress, as is the whole
of this project. The method for analyzing word meanings was developed without any
collaborative input. It undoubtedly needs to be developed further and there is an
urgent need for other researchers to assist in advancing this work. The analyses of
individual words' meanings, even with the tools I have developed, is not easy as the
configurations of factoral constituents involved are not always crystal clear. My
interpretations should not be taken as definitive, but even in its current state this form
of analysis casts new light on word meaning.

11.2 Methodology.
This analysis of word meaning utilizes the factoral structure as set out in Annex 1.
Most of the factors have associated sub-factors (identified by a numerical sub-script
in the formulas). Where there is no sub-script the first, the head factor, applies. The
sub-factors represent specific well-attested ways in which the factors occur in word
meaning, as we have seen. They are natural corollaries of the head factor. Polarity is
prominent. A number of sub-factors have important semantic functions that occur
particularly widely.

Several methodological points need to be made.
1. The analyses, in square brackets, are in two parts, the second enclosed by round
brackets. The first part, following the first square bracket, provides the broadest of
outlines of the meaning in terms of (a) whether the word relates primarily to the
physical world
(as with tree, air, stone), or the human, social world (suggest, talk,
house, person
). Some words have a foot in both camps. (b) Whether the referent is
material or immaterial, i.e. is of a mental nature (believe, remember) or
non-substantial nature (music, love). (c) The word's grammatical function, whether it
is a verb, in which case I use the factor symbol a = action, with a 1–3 superscripts that
indicates the appropriate number of arguments or valencies, plus p or y that
distinguish telic and atelic (punctive and durative) verbs (referring to complete or
ongoing actions). (d) If it is a noun the entry in this position is a p or m3 indicating
that it is a count or a mass noun. (e) Adjectives are indicated by l = display in this
position. (f) Any other grammatical form I identify arbitrarily with a meaningless x.
This use of the letter symbols, a, p, y, m3, l, x are matters of convention that describe
these intrinsic aspects of word meaning that are the foundation of the second part of
the formula. I propose that grammatical role is always integral to the meaning of the
word. It is not imposed by syntax.

2. The next section, the nucleus or kernel of the formula, in round brackets, spells
out the main structure of the meaning in factoral terms, whether quite fully as in the
case of many core and near-core words, or more sparsely as in many other words. In
this part the factors have a rough natural order that is often hard to determine
exactly, with those that are the most salient in forming the meaning coming early and
others that fine-tune the meaning coming later. The meaning is built from the ground
up, as it were. This is just a rough rule-of-thumb at this stage. Sometimes two or
more primitives are sub-bracketed to indicate that they form a composite sub-feature
of the full meaning. There are several kinds of these sub-assemblies that will be
illustrated below.

The analyses take words in particular common contextual senses. A major
complication in this task is that many words are polysemous to various degrees and
even those that are not are frequently malleable to some extent as they need to work
in many contexts. I do not have scope in this paper to explore this important aspect.
It is an outstanding task.

11.3 Spelling word meaning.
I will analyze the meanings of four words quite fully for familiarization purposes and
then provide a number of annotated analyses as further examples.

Duck (n.) is a Type 2 word with a strong perceptual gestalt:

........• belonging to the physical lifeworld
........• materiality
........• a count noun
........• particulateness
........• bodily roundedness
........• animacy
........• display
........• displacement/disparagement

This factoral configuration could relate to almost any animal or bird except that the last
factor is unusual and specific. It represents, arguably, the way people often regard ducks
– as slightly comical and absurd creatures on account of their waddling walk, raucous
vocalisations, beady eyes and wary ways. These are distinctive features that seem
consistent with the last factor. Even then the set of factors is very inconclusive. However
the largely visual gestalt comes to the rescue. The word also has a copious EM. It is
interesting that ‘duck' belongs to the -uck phonaestheme together with buck, chuck,
fuck, luck, muck, suck, tuck
in all of which the displacement/disparagement of /u/ is
active.

The full formula is: *[e, m, p, (p, b, b3, l, u3)]E

Expect is an abstract Type 1 word. Abstract words are resistant to the usual kinds
of compositional analysis. This word has a relatively straightforward meaning.

........• belonging to the human life-world
........• immateriality, (of a mental nature)
........• an act or state involving an actor and an outcome
..........(a transitive verb with two arguments)
........• atelicity
........• relation, dual valency (p:p)
........• is an act of the mind/body
........• is directed externally
........• mental space
........• personal energy
........• emptiness with the power to contain
........• specificity
........• uncertainty
........• fullness of intrinsic value
........• bodily-felt affect
........• positiveness
........• existential value

This is a rich structure. It includes information that is only implicit in the dictionary
definitions such as the affectivities that are not normally expressed in definitions.
The most central constituent is emptiness with the power to contain, a particularly
effective sub-factor. The sub-assembly (p:p) represents by convention the dual
valency or argument structure of the verb. Although it replicates the earlier
superscript of a, it is included here because it is an integral part of the word's
meaning rather than simply part of its description. This sub-assembly is common to
all transitive verbs and to some words of other grammatical foms when valency is
integral to meaning. The next two factors are standard, with variations, to all
transitive verbs. The express basic notions aspects of meaning. The remainder
spell out the idiosyncratic meaning of the verb to the degree that this is possible
factorally. The very useful factor, emptiness with the power to contain, is cental to
the word's meaning. It is surprising that such an abstract word can be analyzed in
such a detailed way by factoral analysis.

The gestalt involved in except is probably a specific proprioceptive (the bodily
disposition of readiness) and interoceptive (affective, the ‘gut feeling' of
anticipation) complex that implies both uncertainty and confidence. EM relates
partly to how the word is used in various contexts.

Here is the full formula:
*[c, m7, a2, y, ((p:p), s2, n5, o2, j2, v5, p, q, v2, s, d, w)]E.

The adjective safe is also Type 1. The dictionary refers to being unharmed, having
escaped from danger, being in sound health, not liable to be lost or harmed, being
relieved, secure, trustworthy, dependable
. The core sense of the word is arguably
captured by free from danger or risk. Here is the factoral structure.

........• belongs to the human lifeworld
........• associated with both materiality
and immateriality
........• descriptive
(adjectival)
........• bodily-felt affect
........• relation, dual valency
........• self-interest
........• physical intimacy
........• satisfaction, comfort
........• fullness of intrinsic value
........• physical or figurative fullness
........• fullness of intrinsic value
........• positiveness
........• existential value

The dual second factor is necessitated by the adjective's association with the
material base of the quality it denotes (here, the subject and the sanctuary) and the
prominence of subjective affectivity in the meaning. The indication of grammatical
role is provided by the third factor. Relation or dual valency here reflects the implied
subject and the implied danger that are integral to the meaning. Physical intimacy
reflects the normal but not necessary association of safety with the intimacy of a
sanctuary and possibly human company. The next four represent the emotional
plenitude of being safe. The last confirms the word's existential import and indicates
its status of membership of the core of the lexicon.

The structure of the meaning of this word is, again surprisingly explicit. But
there is more work here for the gestalt. There is nothing image-like for the gestalt to
work on. It seems to take the form of a distinct but complex proprioceptive sense, a
feeling of the body that builds on the nuclear factors, together with interoception, a
distinctive gut feeling. EM includes associated words like danger, risk, rescue, relief
and schematic examples of achieving safety, being in danger and being safe. The
formula follows:

*[c, m2/m7, l, (s, (p:p), h3, f4, u4, v2, v3, d, w)]E

The preposition on (Type 1) is a closed-class word. As a preposition, it has a much
larger variety of possible contexts and might be expected accordingly, to be highly
abstract. But this is not the case. Here is the set of factors that can be identified:

........• pertaining to the physical lifeworld
........• associated with materiality
........• indeterminate grammaticality (but evident from meaning and context)
........• spatiality
........• display
........• relation, dual valency
........• particularity
........• surface
........• contraction
........• correctness, order and utility
........• existential value

On has a surprisingly competent factoral structure considering its simple
prepositional function. The presence of valency may seem surprising but it simply
reflects the fact that the preposition has two necessary inferences: a subject and a
position. These again are integral to the meaning. Particularity represents specific
position and contraction indicates contiguity with the surface. Correctness, order and
utility
may seem out of place. The rationale for its inclusion is that there is a
paradigmatic sense of correctness about something occupying this position,
particularly in contrast with allied prepositions such as off, under and above that have
connotations of more ‘non-normal' positions. The list is notable for the inclusion of
only two beta factors in contrast to the previous words. The kernel factors, the fourth
to the tenth, build the structure very efficiently, which is just as well because the
gestalt is limited. In formulaic terms the meaning is:

*[c, m2, x, (o, l, (p:p), f, n, u4, w)]E

There is an indistinct but definitive gestalt here that could be either visual or
proprioceptive in nature depending, in part, on whether a thing is seen on, or placed
on a surface.

This is just an introduction to the capacity of the 3-part model and the semantic
factors that is designed to illuminate the nature of word meaning and what we know
(intuitively) when we know the meaning of a word, but it begins to demonstrate the
extraordinary capacity of the factors to build lexical semantic structure. To
supplement this material and to enable the reader to obtain a better grasp of the
model in operation I will set out a larger group of disparate words in the next section.

11.4 More word meaning analyses.
The analyses below provide examples of word meanings of the four types described
above drawn from the early sections of the dictionary including some closed-class
words that are quite challenging. The words are Type 1 unless otherwise indicated. A
considerable proportion of them have more or less abstract meanings. Many of the
words are polysemous to some degree. This implies differences that should be (and
can be) represented in the formulas. Illustrating this is not practicable here. I have
tried to use a common sense of the words.

I submit that the examples show that the model is highly successful in clarifying
how word meaning works. This seems to be simply a consequence of the nature and
function of the factors, their primal lineage, and their ability to apply across the
lexicon and human concepts in association with gestalts and EM. The factors provide
a natural fit for the description of word meaning at the structural level. I must,
however, repeat that my analyses remain provisional until other researchers work
through the process I have used.

able – *[c, m2/7, l, ((p:p), j2, g3, v3, q2, d1, w)] E
As in ‘He is able to work'. The gestalt has a mainly proprioceptive form, a bodily sense of readiness for
a task. The second, dual, factor states that the word involves physical and mental capacities (m2 and
m7) and the third that it is an adjective (l). The sub-assembly indicates relation or the dual valency of
the word, its inegral linkages with a verb and a prepositional phrase. The dominant nuclear factor is
personal and physical energy (j2) that is coupled with effectiveness and fullness to express a
substantial part of the meaning.

about – *[c, m7, x, ((p:p), g3, l, q2, w)E
This is a polysemous word. I use it here in the sense it has in ‘This book is about Spain'. In this sense
the bracketed sub-assembly indicates that the word has two valencies (book and Spain). The meaning
is effectively conveyed by g3, paricularly in its sense of affordance, combined with l. Aboutness is a
relation in which content or meaning is provided.

above – *[e, m2, x, ((p:p), o, l, u, d1, w)] E
This word poses a challenge to factoral analysis. The sub-bracketed p's indicate again a valent
relationship: subject and position (‘the dot is above the line'). As there is no factor for high/low I have
used u, displacement, to represent a departure from the ‘normality' of on, and d1, positiveness, to
convey the normal positiveness often associated with being above. The lack of an elevation factor is
indicative the extreme primitiveness of the factors.

accident – *[c/e, m2, p, (u, l, k, d2)] E
A near-core word. Much of the semantic burden is carried by u1/2, and d2, displacement, wrongness;
and negativeness. The k represents the shock of suddenness and unexpectedness. The gestalt
seems to have elements of perception (of a typical accident) and a ‘gut feeling'.

acid (sour) – *[c, m2, l, (s3, k, d2)] W E
A Type 3 (qualic) word. The structure could fit numerous such words but the gustatory gestalt is
precise.

act (v.) – *[c, m2, a1, p, (s2, n2, r, a, p, g3, l, o, s, w)]
This is our first verb, an intransitive verb as indicated by both a1 and n2. The first two factors of the
kernel are common to all intransitive verbs, the n2 expressing intransitivity in factoral speak, the sense
of confinement of the action to the subject. This is quite an effective description of a word with a simple
sense that largely coincides with the factor a. The g3 (affordance) represents the implied consequence
and the s, the vague affective sense that action is to be preferred over stasis.

add – *[c, m2, a2, p, ((p:p), s2, n5, r, a2, l, v3, w)] E
A transitive verb. The a2 factor, enlargement, is the natural fulcrum here as its meaning is close to that
of the word. The n5 conveys the sense of an action that has an external effect. R indicates a degree of
effort.

aeroplane – *[e, m, p, (p, m, l, o, r4)] E
A Type 2 word. The factors produce only a limited structure but the visual gestalt is strong and EM
very extensive. The final factor is intended to convey the impression of a ‘plane flying effortlessly. The
gestalt and EM do almost all the work.

air – *[e, m, m3, (k4, (o, a2), h, g2, d1, w)] E
A core Type 2 mass (m3) noun. Unsurprisingly, the factors cannot provide a satisfactory description,
despite trying hard, because ‘air' is so, well, airy. The factors represent key qualities to an unusual
degree for a Type 2 word but the gestalt and EM carry much of the semantic load. The gestalt is
probably associated with the vastness of the atmosphere and breathing, a challenging semantic
scope.

and – *[c, m2, x, ((p:p), a2, m3, w)]E
This and the following two words have considerable common semantic content as expressed by a2 but
the kernel of the formulas distinguishes them quite effectively. The sub-assembly simply indicates that
the word operates as a linkage between two passages. The nature of the limkage is specified by the
following factors.

all – *[e, m2, x, (a2, p, g, v3, l, s, w)] E
The p indicates the finiteness of all.

also – *[c, m2, x, (p:p), a2, t3, g, s, w]E
The t3 expresses the integral sense of pointed reference that is absent from the previous two words.

although – *[c, m7, x, ((p:p), q, v5, u4, w)] E
This may seem an impossible word to analyse in this manner. The sub-assembly represents the
inherent valency that links with the word's complements. We will see that v5, emptiness that is a
capacity to contain
is a very useful sub-factor. The word implies a void that is to be filled.

ankle – *[c, m, p, (p, s4, w)] E
Type 2. The formula is totally uninformative. The gestalt and EM fill the gap.

apple – *[e, m, p, (p, b2, i, n, l, g3, w)] E
A typical Type 2 word with a strong gestalt. The n relates to the graspability of a fruit. The formula
could apply to any fruit.

as – *[c/e, m7, x, ((p:p), l, t3, q2, w)] E
In the manner of; in the same way that. ‘Good as gold'. This simple formula seems to produce an
appropriate semantic structure with t3, pointing, deixis, as its fulcrum.

ask – *[c, m7, a3, p, ((p:p:p), s2, n5, n4, l, q2, v5, w)] E
A simple and effective formula for a cognitively simple concept. N4 = expression. The verb has three
valencies or arguments, an asker, the person asked and the question. Q2 indicates the definitiveness
of the question and v5 the expectation of an answer.

bad – *[c/e, m7, l, (d2, u2, v3, l, w)] E
The description is succinct because most of the meaning is expressed by the first and second factor in
the kernel, negativeness that is augmented by u, displacement and v3, fullness (acting as an
intensifier).

badminton – *[c, m2, p, (j2, f4, j, u2)] E
Type 4. The first kernel factor is intended to convey personal commitment as required by a sport, the
second, sociality and the last two, enjoyment and satisfaction. This could apply to most games.

be – *[c/e, m2, a1, m3, ((p:p), r2, n2, i3, q2, (p, l, o), v3, w)] E
A central word in English. As in ‘What can it be?' The same formula applies to its other forms, am, is,
are, was, were
with some small variatiojns. This is a notoriously difficult word to define but the formula
is surprisingly effective. The r2 designates being as a process. The i3 (identity) provides much of the
semantic thrust. The n2 indicates that this is a centripetal, semantically self-referring verb (duplicating
a1). It nonetheless has a binary valency as is clear from an answer to the above question. ‘It is rump
steak'.

bed – *[c, m, p, (p, f2, l, g3, u4, w)] E
Type 2. Similar to apple. This is a bare structure for the gestalt.

before – *[e, yt, x ((p:p), (y2, n, p), w)] E
This is another formidable challenge. I have adopted yt as a convention for marking temporal words.
The second sub-assembly can be paraphrased as ‘the period before the initiation of an event'.

behave – *[c, m2, a1, y, ((p:p), s2, n2, h4, (h3, g3), l, o, s, w)] E
This is an intransitive verb as stipulated by the a1 and n2. The next three factors present much of the
meaning. The contiguous l and o occur together commonly. Here they indicate the overt manifestation
of behavior.

behind – *[e, m2, x, ((p:p), (o, n), u, l3, w)] E
As with ‘before' the kernel factors form an effective analogue of the word's sense. The o, n indicates a
constricted space. The word is the obverse of ‘before' in its spatial sense where the kernel is (f, o, n, l,
u4, w).

believe – *[c, m7, a2, y, ((p:p), s2, n5, (o2, i4), l, g3, v2, u4, s, w)] E
This is another effective structure for a highly abstract verb.

below – *[e, m2, x, ((p:p), o, (f, u, l3, d2, w)] E
This is the converse of above.

big – *[e/c, m2, x, (a2, b, l, o, d1/2, w)] E
B here provides the sense of bulkiness that is often integral to the word's meaning. The two d's are
indicative of the positive and negative affectivities often associated with the word in different contexts.

bird – *[c, m, p, (p, b, b3, l, o)] E
This Type 2 word illustrates again the dependence of this group in their explicit gestalts and EM.

blanket – *[c, m2, p, (p, f2, f4, s3, g3, u4)] E
This is an unusually explicit formula for a Type 2 word. The factors represent some of the typical
qualities of a blanket quite effectively.

blue – *[e, m2, l, (s3, f, l, k)] E
Type 3. The perceptual gestalt carries the full semantic weight. EM provides information, inter alia, on
where various shades of blue are found. K is indicative of the idiosyncracy of the colour.

body – *[c, m, p, (h4, p, b, h3, g3, l, o, w)]E
I am using the word as referring to the human body. The gestalt has elements of proprioception
and perception.

both – *[e, m2, x, ((p:p), t3, l, o, w)] E
The sub-bracketed p's indicate the duality at the centre of the meaning. The combination of l and o
simply reinforces the deixis (indication) of t3.

bowl (n.) – [c, m, p, (b2, v5, l, o, w)] E
The first two kernel factors represent, in effect, the outside and inside of the bowl respectively. The
visual and tactile gestalt carries considerable semantic weight.

breast – *[c, m, p, (s4, p, (b, g3), v2, w)] E
The robust proprioceptive, visual and tactile gestalt is well supported by the factors.

but – *[c, m7, x, ((p:p), u, q, v5, w)] E
The first three kernel factors are remarkably effective in sketching the word's sense with
considerable weight on the third. The relation sub-assembly is necessary to indicate the two
complements of the word.

by – *[e, m2, x,((p:p), o, n, f4, l)] E
Taking the word in its spatial sense, the f4 (contiguity in effect) provides much of the content.

can – *[c, m7, a2, y, ( s2, n5, (p:p), o2, j2, (a, g3,), v3, q2, w)]
This is a modal auxillary verb that means ‘to know how to'. Its meaning is close to that of
‘able'. The g3 has the sense of ‘affordance' that expresses the ‘how'. Alternatively, this could
be expressed by v5.

care (for) – *[c, m2, a2, y, ((p:p), s2, n5, j2, f4, t, n, g2, v2, s, d1, w)]E
The f4 represents the nurturing environment, the n conveys the focus of attention that is involved in
caring and g2 the all-important sense of generosity.

carry – *[c, m2, a2, p, ((p:p), s2, n5, h4, (h, n), (r, m3), v3, w)] E
The h4 indicates that a person is doing the carrying. The second sub-bracketed factors represent the
close holding that is normally implied and the third, the sense of conveying.

catch – *[c/e, m2, a2, p, ((p:p), s2, n5, (p, n4>n, h),u4, d1, s, w]E
The > means ‘becomes'. Detachment becomes capture and possession.

chair – *[c, m, p, (p, l, g2, u4, w)]E
Type 2. Similar to ‘bed' in its non-specificity.

cause (v.) – *[c/e, m2, a2, p, ((p:p), r2, n5, r, g3, q2, v3, l, d1, s, w)]E
The main semantic thrust of this important word comes from g3 and q2.

change – *[c/e, m2, a2, p, ((p:p), r2, n5, r, (p>p), g3, q2, v3, l, d1, w)]E
A state becomes another state as the sub-assembly indicates. These last two words are quite closely
related. To cause is to change and to change is to be caused to do so.

clean (adj.) – *[c, m2, l, (f, l, u4, s, d1, w)]E
Surface has a pivotal role here.

come – *[c/e, m2, a1, p, (r2, n2, h4, r, m3, (y>n), l, o, w)]E
In the sense of ‘here he comes'. The sub-assembly conveys the sense of diminishing
distance. In ‘go' the sub-assembly would be (n>y).

consider – *[c, m7, a2, p, ((p:p), s2, n5, j2, o2, (g3, q), s, w)]E
The second sub-assembly converys the sense of deliberation without a firm outcome.

11.5 Summary and comment.
I must stress again that there is nothing definitive about these analyses because this
kind of task is so new and has not been subject to independent scrutiny. The finer
points of interpretation are often difficult to make. Nonetheless we can draw some
interim conclusions from the results:

1.....These intuitive analyses, have produced some fairly clear-cut results that can be
.......tested by other researchers.

2.....By and large the validity and veracity of the factors is supported by the work they
.......do in the analysis. In particular, they are remarkably effective in delineating the
.......structures of the meanings of the important Type 1 core words and near-core
.......words including those with abstract meanings and closed-class words. This is
.......perhaps their most striking achievement. No other lexical semantic analysis has
.......demonstrated this capability.

3.....The affective, beta factors are demonstrated to have a critical role in extending
.......word meaning beyond the description of physical entities. In their absence
.......language would be a mere shadow of its present form. They have a central
.......function in many abstract and semi-abstract word meanings.

4.....Gestalts lack the independent authorization that the factors possess in articulatory
.......iconicity, but the factoral configurations of structure themselves provide a
.......warranty for the gestalts they underpin. The concept of semantic gestalts is also
.......supported by the way their full form can arguably be ‘felt' in each of the words,
.......including closed class words like but and by, in ways that are illuminated by
.......knowledge of the factors.

5.....Encyclopedic meaning does not need any further justification as its presence is
.......generally clear in the specialist words and the specific uses of all words where it
.......needs to come into play in fine-tuning meaning to context.

6.....The concept of the four types of word meaning is supported by the distinctions
.......between the analyses of each type.

7.....An important aspect of word meaning that has not been addressed in this paper is
.......its linkage with the ways in which words combine in phrases and sentences. The
.......degree to which word meanings contain syntax to some degree is a matter for
.......further exploration (14.4).

I submit that this model, with the semantic factors at its centre, provides a
more empirically sound and a more effective and satisfactory description of
word meaning than any other approach in the literature. Its outstanding
achievement is its ability to analyze the meanings of all types of words in any
language with the same economical analytical equipment.

11.6 The empirical reliability of the factors.
For the first time, it seems, we have an objective means of looking inside word
meaning with some confidence. We are better placed than ever before to determine
the ‘dark matter' of word and concept meaning and cognition. We can now provide an
answer to the question I posed at the beginning of this paper: What do we know
when we know the meaning of a word? The model indicates that there are three parts
to the answer but the factors have a central function. We do not ‘know' the factors in
a conscious sense, but as dimensions they are present and can be intuited with
some difficulty and rather vaguely without knowledge of their identity, but with
considerable confidence with this knowledge. This gives the mental content of word
meaning a transparency it has never had before. This is one of the main
achievements of this project.

My expectation is that the factors will help to place lexical semantics on a
footing that approaches that of the physical sciences in terms of objective explanation
of how word and concept meaning operates despite the totally different nature of
lexical entities and the objects of physics or chemistry. What is required to more
firmly establish this position for word meaning is neurological confirmation that the
factors can be reliably correlated, in a systematic manner, with specific locations in
the brain that are operative when words are understood. Neuroscience is beginning
to make progress in this exceedingly difficult task but without the assistance that the
factors may be able to provide. Much remains to be done (see Section 14). I believe
the factors and the 3-part model have the potential to facilitate this research.
However, they are unlikely to be utilized unless linguists and psychologists come to
provide some authoritative support for what might otherwise be regarded as a
maverick proposal. This is a shortcoming that this paper is seeking to resolve.

There is, however, another possible route for verifying the existence of the factors as
psychologically real aspects of word meaning. If they exist in the meanings of most
core words as I propose, it should be feasible to devise a project along the lines of
the Feature Norms of McRea and associates (see 12.9), but taking account of the
type of semantic feature that the semantic factors represent, as well as the other
elements of the 3-part model. This would involve using a sample of core words and
asking a group of suitable subjects to break their meanings down into semantic
features. The subjects would be given guidance as to the desirable level of
abstraction and generality of the target semantic features but not being given the
factors themselves. For example the 60 or so primes of Anna Wierzbicka's and Cliff
Goddard's Natural Semantic Metalanguage (NSM) as discussed in 12.3 could be
used. They have an exceedingly abstract and generalized character but each one
can be analyzed factorally given knowledge of the factors. Without that, but with only
broad guidance, subjects should be capable of producing a small number of features
that could then be ‘normalized'. Using ‘somebody' as an example it would be
interesting to see how close the result would be to the analysis that can be made
from knowledge of the factors – belongs to the human domain, materiality,
particularity (is a count noun); personhood, generality, uncertainty, fullness of positive
value, existentiality
. There are some procedural variations and refinements that could
be used to achieve a satisfactory result. The project would require much more
application from subjects than was obtained in the original feature norm project. This
would require a degree of carefully targeted guidance so as to achieve the
appropriate level of abstraction that would be supplemented by gestalts and EM. The
results should be of value to researchers such as neurophysiologists who are
investigating the representation of word meaning in the brain.


annotated contents




12. Theories of word meaning.

Although a number of significant approaches to word meaning are still on the table,
lexical semantics is arguably the most under-developed major sub-discipline of
linguistics. There seem to be several reasons for this. First, the intuitive nature of
word meaning makes it a difficult field in which to achieve concrete, let alone
scientifically acceptable, results. This inhibits researchers from exploring the subject
when more tractable aspects of language such as phonology, corpus linguistics,
pragmatics, sociolinguistics and psycholinguistics seem to have more potential.

Second, there are doubts about investigating word meaning because words are just
elements in the larger constructions of language that themselves yield information
about words' meanings. Generative grammar is well-known, if not notorious, for its
belief that syntax and grammar subsume semantics to a large degree. Third, despite
the endeavours of dedicated researchers such as Fodor, Fillmore, Jackendoff,
Pustejovsky, Barsalou and others there has been little agreement on the actual
nature of word meaning. There has been no gradual consolidation of basic principles
such as has occurred in many other areas of linguistics. Some authorities are
probably simply willing to bide their time and wait for answers to be provided from the
neurophysiological investigation of word meaning. However, neuroscience itself may
be seriously impeded by the failure of linguistics to provide it with some leads in this
exceedingly complex task.

Within linguistics a number of theories of word meaning and attempts to throw
light on its nature have emerged over the last fifty years since interest in the subject
was renewed in association with the development of generative grammar. A variety
of approaches are still in circulation. In this section I will outline some of these
theories and I will assess them in the light of the model set out in Section 9. This
review will be highly critical because none of the other approaches appears to have
as sound an empirical basis as the 3-part model or to be able to analyse the
meanings of various types of words as effectively. On the other hand, some of these
theories contain elements or aspirations that are highly compatible with my approach.
Considerations of space preclude anything but a very brief description and evaluation
of a selection of these theories.

The theories discussed are: 1. Early lexical decompositional theories. 2. Lexical
Atomism (Fodor). 3. The Natural Semantic Metalanguage (Wierzbicka and Goddard).
4. Conceptual Semantics (Jackendoff). 5. The Generative Lexicon Theory
(Pustejovsky). 6. Cognitive Semantics (Lakoff and Johnson). 7. Prototype Theory. 8.
The Perceptual Symbol System (Barsalou). 9. Feature norms (McRea and
colleagues). 10. Neurophysiological investigations.

12.1 The early decompositional approach. The term decompositional denotes an
approach to word meaning that emerged in the 1960s (Katz and Fodor, Bierwisch,
Weinrich and others). Like the three-part model this approach assumed that word
meanings are complex, that is, they have an internal structure of semantic features
(sometimes called semantic markers, semantic primitives or primes) that can be
identified intuitively. It became evident early on that a satisfactory set of semantic
features, one that would permit an effective description of word meaning, would need
to meet specific criteria such as primitiveness and abstractness, finiteness and
universality. Bierwisch 1967 wrote perceptively:

..........There are good reasons to believe that the semantic markers [features] in an
..........adequate description of a natural language do not represent properties of the
..........surrounding world in the broadest sense, but rather certain deep seated, innate
..........properties of the human organism and the perceptual apparatus, properties which
..........determine the way in which the universe is conceived, adapted and worked on.

However, the practice did not match the theory as researchers found it impossible to
identify features that met these requirements in convincing ways. The early attempts
developed from an analogy with a then prominent phonological approach to the
description of vocal sounds using a small set of features of the articulation of sounds
(distinctive features) and describing individual sounds in a binary (+/–) fashion
according to the presence or absence of specific features. Using a similar format
semanticists identified binary semantic features like male/female, alive/dead,
open/shut, married/unmarried
and applied them to a narrow lexical field, mainly
nouns. This approach was eventually abandoned as its limitations were recognized.

12.2 Fodor's lexical atomism. The concept of word meaning developed by the
American philosopher Jerry Fodor (1975), a major contributor to linguistic thought at
the philosophical level, is diametrically opposed to decomposition and most of the
approaches to word meaning described here. Fodor regards word meanings as
atomic and irreducible. Decomposition is abjured. Meanings have no internal
structure and so are irreducible to semantic features of any kind because, he holds,
there is a direct correspondence between words and the mental particulars they
stand for. Words simply denote. Word meanings are inherently primitive. He holds
that there is a nomic or lawful relationship between the senses of words and the
entities that they denote. He also famously characterizes word meaning as innate.
This is a highly controversial position because it implies that the meanings of words
such as democracy, scissors, computer, quasar and the new words to be produced
by the next generation are unlearned and innate, a result that is generally regarded
as absurd.

Although the innateness that Fodor proposes is counterintuitive his belief that
words simply denote without the influence of internal constituents has some intuitive
appeal because we know what words mean in an instant and it is far from obvious
what semantic constituents, if any, might be involved. This is a conundrum that I am
seeking to explain by using gestalts to represent this immediate meaning. Fodor's is
a philosopher's stance that he has argued with great vigor for decades. His position
requires serious consideration despite its apparently maverick qualities.

12.3 The Natural Semantic Metalanguage or (NSM) has been developed by
Australian linguists Anna Wierzbicka, Cliff Goddard and colleagues over the past four
decades. The NSM does not seek to identify primitive word-internal constituents such
as have been sought by other researchers and are the subject of this paper. Rather it
claims to identify a set of semantic primitives or primes in terms of a minimal set of
linguistic concepts, words, with meanings that are universal and irreducible and that
that can be used to describe the meanings of all words in any language. These are
held to form a truly scientific basis for lexical semantics. The latter point has been
heavily emphasized as a unique feature of this approach. Exhaustive investigations
of a substantial number of languages from around the world have resulted in a set of
some 60 primes that are ordinary language words such as I, you, someone, people,
body, kind, part, this, some, all, good, bad, big, small
. These words are used to
generate descriptions of the meaning of words that are wordy (!), sometimes very
wordy. Part of the rationale of NSM is that such descriptions avoid the circularity of
dictionary definitions that often use words reciprocally to describe one another. With
its tiny vocabulary of words that cannot (supposedly) be broken down into still more
basic units NSM avoids circularity. Whatever its merits the NSM does not address the
issue that most investigators face: forming a psychologically and neurally realistic
account of how we intuit word meanings.

12.4 In his Conceptual Semantics, prominent American linguist Ray Jackendoff 2002
has developed a comprehensive framework for the study of language including word
meaning. Jackendoff has investigated word meaning as intensively as any other
contemporary linguist. Semantic features of several types are an important part of his
system but he concedes that the identification of a satisfactory core of primitives is
still a distant goal. In particular he notes four issues: the difficulty in determining the
optimal level of primitiveness; no set of primitives is likely to provide a complete
description of a word's meaning, leaving ‘the problem of completers'; primitives seem
unlikely to account for the fine distinctions in the meaning of related words such as
break, crumble and shatter; and primitives of a perceptual nature are inadequate for
abstract words, leaving a need for abstract primitives. He observes that ‘such facts,
which we confront at every turn, threaten to undermine the prospects of completely
decomposing words into primitives that are descriptively useful and that have some
plausibility for innateness' (2002, p.338). Jackendoff avoids calling them primitives
and adopts other descriptions that reflect the lack of true primitiveness.

He holds that lexical meaning has two main elements, spatial and conceptual
structure. The first ‘is concerned with encoding the spatial understanding of the
physical world' and the second ‘is a hierarchical arrangement built out of discrete
features and functions; it encodes such aspects of understanding as category
membership (taxonomy) and predicate-argument structure'. Spatial structure is not
‘imagistic' but it has the form of ‘abstract image-schemas….. from which images can
be generated'. He notes that spatial structure is useful in dealing with the ‘problem of
completers'. Jackendoff's main interest is in words' conceptual structures and
particularly their role in the construction of phrases and sentences. He consequently
focuses largely on verb meaning with its syntactic argument structure. He identifies
several types of semantic features. First, core functions (such as be, stay, go,
extension, orientation, inchoative
(the beginning of a state), second, ontological
categories
(such as object, event, path, place, state) and third, semantic field
features
that determine the character of arguments. These include spatial,
possession, ascription, scheduling
. With this open-ended mélange of features he
seeks to describe the mechanics of phrase construction. This is an ambitious
exercise and he recognizes that his work is just the beginning. “There are just so
many goddamned words, and so many parts to them". [p. 377]

Jackendoff observes at the conclusion of his chapter on lexical semantics: ‘It
should be clear by now that the generalizations of word meaning cannot be studied
without a theory of lexical decomposition. The kind of decomposition required is a
richly textured system whose subtleties we are only beginning to appreciate……It
remains to be seen whether all this richness eventually boils down to a system built
from primitives, or if not, what the alternative might be'.

12.5 The Generative Lexicon Theory. The originator of this theory, American linguist
James Pustejovsky, operates under the rubric of computational linguistics, an
interdisciplinary field that deals with the statistical or rule-based modeling of natural
language from an information-processing perspective. It originated from the needs of
machine-based translation for an appropriate understanding of the semantics of
words.

One of Pustejovsky's aims is to explain the creativeness of language, its
infinite productivity in using words of apparently fixed meaning by combining them in
sentences in flexible ways – a laudable task. He also seeks to understand the
mechanisms behind the continual extension of words' meanings. He places
ambiguity and polysemy at the centre of the creative capacity of language as a
means of facilitating the co-compositionality of words, the way in which they
mutually accommodate one another in creating meaningful phrases and sentences. A
central task is analyzing the meanings of words so as to identify semantic
constituents that allow their combination in specific ways. He holds that cocompositionality
is not a function of the static meanings of individual words but is
contingent on compatible constituents in the meanings of combined words. The
theory attempts to spread the semantic load across all constituents of an utterance.
He proposes that the semantic constituency structure of words can change to
accommodate the needs of contexts.

Pustejovsky holds that ‘a semantic language is defined by the rules generating
the structures for expressions rather than by the vocabulary of primitives itself.'
Generative rules and devices are prescribed to account for language's
co-compositionality and creativity. His theory is distinguished by the innovative way he
analyses word meanings. He specifies three components in the structure of word
meaning: argument structure, event structure (type of event and any event
substructure) and qualia structure. One of his main innovations is the concept of the
qualic structure that seeks to provide a succinct description of what a word denotes.
This culminates in a set of semantic features that ostensibly lack the primitiveness,
finiteness, universality and applicability to all types of words that are desirable in a
set of semantic features. This theory is original and it addresses important issues. It
has proved to be quite influential.

12.6 Cognitive Semantics is a branch of Cognitive Linguistics (CL), a prominent field
of linguistic research that emerged some four decades ago and that has achieved
considerable prominence. CL arose out of disenchantment with Chomskyan
generative grammar in the 1960 and 70s, developing as a reaction to its narrow focus
on the syntactical and grammatical aspects of language. Some central postulates of
CL are that: semantics is central to the study of language; the human mind can be
studied through language; language is integral to cognition; it is not structured
arbitrarily but is motivated and grounded in experience; word meaning is derived from
concepts; and there are no clear boundaries between linguistic knowledge and
encyclopedic meaning.

The concept of embodiment has become prominent in recent linguistic thought,
especially that of two major figures of Cognitive Semantics, the linguist George Lakoff
and philosopher Mark Johnson. The term embodiment of meaning, refers to the view
that, contrary to the popular perception of meaning as something abstract and purely
mental, it is profoundly influenced by the bodiliness of those in whom meaning
occurs. ‘Meaning reaches deep down into our corporeal encounter with our
environment' (Johnson 2007 p25). Embodiment theory recognizes an aspect of
language that had been neglected. Meaning, and especially word meaning, must
reflect and represent the relationship of the human body as the repository of
meaning, and its human and physical environment that is the subject of meaning.
The term highlights the numerous ways in which aspects of bodily awareness and
movement are important in cognition and word meaning. The idea of the embodiment
of conceptual structure and word meaning has been widely accepted over the last
thirty years by psychologists and linguists as a response to the virtual exclusion of
the body in much previous philosophy of language and of mind that focused on
language as a mental or social phenomenon.

Lakoff and Johnson also attribute a key position in cognition to image schemas
which Johnson (2007) defines as ‘dynamic recurring patterns of organismenvironment
interaction' that ‘are revealed in our basic sensori-motor experience'. (p
136). They are ‘recurrent patterns of sensorimotor experience; ‘image'-like in
preserving the topological structure of the perceptual whole; both ‘bodily' and
‘mental'; predicated on interaction with the wider environment; realized as activation
patterns in topological neural maps; and structures that link sensorimotor experience
to conceptualization and language'. He cites schemas such as container, verticality,
compulsive force, scalarity, source-path-goal
as representing fundamental features
that can be extended metaphorically into numerous areas of word meaning and
usage including the representation of abstract concepts. Johnson emphasizes the
role of image schemas in ‘conceptualizing and reasoning about abstract domains'.
He discusses embodiment in relation to concepts but he does not explore its function
in the decomposition of the meanings of words. Nonetheless, there are some
significant parallels between the way Johnson describes image schemas (p. 144)
and the semantic factors.

Cognitive Linguistics has made some useful contributions to semantics but it
has not produced a viable account of word meaning.

12.7 Prototype Theory was developed in the 1970s by Eleanor Rosch (1978) and
other American psychologists (Rosch (1978) and Rosch and Marvis (1975)). The
theory sought to account for the widely recognized fuzziness of categories and word
meanings by using prototypes that display the typical features of members of a
variable category such as bird or vegetable. This approach contrasts with the
classical Aristotelian view that categories and word meanings are checklists of
necessary and sufficient conditions. Prototypes are generally conceived as
categories that are rooted in experience – they are learned concepts of the world that
are grounded in shared, embodied experience. They are intuited as coherent,
meaningful wholes that consist of a number of associated features, described in
words that combine in a psychologically realistic way. Decomposition is thus inherent
in this approach. Prototypes highlight the idiosyncratic, descriptive element of word
meaning but, contrary to the imagistic connotation of the term, most authorities
regard them as being composed of lists of surface-descriptive features (such as for
‘bird', has feathers, has a beak, can fly). Features can be given statistical weightings.
Prototypes were identified by psychological investigations with respect to natural and
artifactual things such as birds, fruit and furniture. The theory is less satisfactory for
analyzing the many other types of words that constitute the bulk of the lexicon.

12.8 Perceptual symbols. The American psychologist Laurence Barsalou's
Perceptual Symbol System offers a quite different approach from most of those we
have been discussing. It focuses on the sensory aspect of concept and word
meaning. After a long period of neglect this aspect was given prominence in the work
of a number of authorities late in the twentieth century. I will describe Barsalou's
influential system in more detail because its parallels with the 3-part model are
striking. The basic principle of his approach is that ‘a common representational
system underlies perception and cognition'. Barsalou (1999) emphasizes the modal
(sensory) form of cognition. Word meaning has a sensory basis. According to this
view the meaning of ‘kick', has a dominant kinaesthetic component, a ‘kick' within the
brain, as it were. He also proposes that in every use of a concept or word, the
meaning is not simply retrieved from long-term memory. It is reactivated or
simulated in the region of the brain where the original neural formation of an
experience occurred. Such simulations are a key feature of Barsalou's approach.
‘Simulation is a unifying computational principle throughout the brain' (2007). In the
use of words their meanings are simulated or re-enacted in somatotopic (bodyrelated)
regions of the brain. Meaning is thus firmly grounded in non-linguistic
physical cognition of the world.

Barsalou proposes several key attributes of his perceptual symbols. They are (i)
sensorimotor neural representations; (ii) they operate below the level of
consciousness; (iii) they are analogical in that they represent necessary features of
entities, as opposed to being amodal, abstract and arbitrary; (iv) they are schematic
and compositional rather than holistically pictorial; (v) they are dynamic in their ability
to generate mental forms; (vi) they are multimodal as they can occur in any of the
four sensory modalities as well as in proprioception and introspection
(representational states, cognitive operation and emotional states) and (vii) they can
combine modalities.

Barsalou and his collaborators have argued their case persistently and
forcefully and assembled some empirical support for their system but he concedes
that much further work is required to provide a sound empirical base for the theory.
Barsalou's perceptual symbols are one of the most significant recent proposals in the
field. However, some of Barsalouʼs main conclusions are being questioned and
amodal cognition still has a strong following in some quarters.

12.9 Feature norms.
A number of researchers have recently used a different approach to produce sets of
lexical semantic features that are eminently obvious and highly descriptive, what I
have called surface-descriptive features (SDFs) (McRea et al 2005, Cree at al
2006, Vigliocco et al 2004, Moss et al 2007). Such features have been obtained for
specific words by a procedure that uses groups of subjects under carefully controlled
conditions, to generate features associated with words as they come to mind. The
resulting lists of features are then coordinated to produce so-called feature norms,
sets of weighted features that provide information about the content of concepts that
are taken to be representations of the meanings of the words. For example, such a
process investigating the meaning of ‘dog' might produce sets of features like: is an
animal, has four legs, is furry, barks, wags its tail, can bite, makes a good pet
.
Feature norms are effective in that there is no doubt as to what is being described.
They operate like crude dictionary definitions, but they have obvious weaknesses that
I will outline shortly. Sets of feature norms have been used in neuropathology for
interpreting the neural causes of specific cognitive deficits in people with brains
damaged by illness or accidents. A developing literature testifies to the utility of this
type of semantic feature and its impressive credentials but it has some severe
limitations as an explanation of word meaning.

12.10 Comment on the theories.
My comments on these approaches to word meaning come from the purported
vantage point of the 3-part model with its firm empirical basis in the lexicon. I will
comment from the perspective of the light shed by the similarities and differences
between the approaches sketched above and my model. My conclusions are highly
critical as they rest on the semantic factors' testable presence in word meaning.
Other approaches generally lack a defensible empirical foundation.

The early decompositionalists such as Fodor and Katz, Wienrich and Bierwisch,
who insisted on the complexity of lexical meaning arising from the presence of
semantic features, established an important principle for much of the discussion on
the nature of word meaning over the following half century, but failed to identify
features that were sufficiently abstract to provide a workable kind of analysis.

Fodor's atomistic account of word meaning is discounted by most authorities as
it is clearly untenable in its original form of the innateness of word meaning. However,
if an innate structure is assumed at a more primitive (feature?) level, as Fodor seems
to have come to concede, his approach becomes much more tenable. It is consistent
with the nature of the semantic factors as necessary constituents of word meaning in
my scenario. The ability to form sensory-motor and experiential gestalts with factoral
structure has an equivalent innateness and a superficial semblance of atomicity.
Word meanings can be conceived as atoms with sub-atomic structure. In
combination these may well enhance the non-linguistic Language of Thought that
Fodor has propounded, the semantic factors being its alphabet! There is a deep
compatibility here if Fodor's combination of purely lexical atomicity and innateness is
abandoned. The concept of a non-linguistic Language of Thought that under-pins
expression in natural languages and in deliberative thought is difficult to avoid and
has been accepted by many theorists in one form or another.

The Natural Semantic Metalanguage of Anna Wierzbicka and colleagues has
little in common with the 3-part model as the primality of its primitives is of an entirely
different order and it does not address the content of word meaning except in a
lexicological, dictionary-making way.

Conceptual Theory. Jackendoff has been a persistent investigator in his
attempts to identify a satisfactory kind of semantic feature but by his own admission
his progress over three decades has been modest and inconclusive. His insights into
word meaning are substantial but it is doubtful whether his tentative conclusions
provide a useful basis for future investigation. The lack of a robustly empirical
analysis of lexical content in his work is a serious shortcoming. His approach has
been influential but the 3-part model seems to address many of the issues he
identifies more effectively. His spatial structure has much in common with gestalts
while his conceptual structure is, I believe, handled more elegantly by the 3-part
model.

Pustejovsky has developed a novel and interesting approach that is complementary
to Jackendoff's but his work also lacks a sound empirical foundation.
The main contributions of Cognitive Linguistics in this field have been those of
Lakoff and Johnson with their work on embodiment and image schemas. Johnson, in
particular in his 2007 book, with its discussion on organism-environmental interaction
as a precursor of concept meaning, embodiment and image-schemas, has made an
important contribution. These notions are highly consistent with aspects of the 3-part
model and the discussion in Section 10. Johnson's book attempts to combine
contributions from the American pragmatic philosophers with those from
neuroscience but ultimately his conclusions lack empirical clout.

Prototype Theory made a valuable contribution to the development of thinking
on concept and word meaning. One of its limitations, similar to some other theories,
is that it works satisfactorily with concrete entities but is of little value with most other
words.The notion of mental prototypes has something in common with the gestalts of
the 3-part model but they transcend its limitations of application.

Barsalou's Perceptual Symbol System (1999 and other papers) has some
strong parallels with the 3-part model. He has persistently opposed the concept of a
purely amodal (not based on sensory features) description of word meaning
championed in particular by Pylyshyn (1984) and Fodor. The properties that Barsalou
attributes to his perceptual symbols – their sensory-motor neural basis, analogical
character, schematic nature and compositional scope – are consistent with my
model. A major difference is that Barsalou' system recognized and utilized the
sensory-motor properties of word meanings without due recognition of the affective
aspects represented by the beta factors. His perceptual symbols are gestalts devoid
of the intrinsic affective content that is a vital component of the 3-part model. This
has been the source of his system's inability to cope with non-concrete words without
the use of cumbersome context-based explanations in contrast with the facility of the
3-part model in dealing with this type of word. More recently, however, Barsalou and
his colleagues have worked on a range of aspects of emotion and have produced
conclusions that are consistent with the role of affects in the 3-part model (Wilson-
Mendenhall, Barrett and Barsalou (2013) and other papers). This work reflects the
growingcurrent research into aspects of emotion.

Feature norms. I will discuss these in a little more detail because they may
appear to be a substitute for the semantic factors. Surface-descriptive feature (SDF)
norms describe certain types of word meaning (mainly basic-level and other nouns)
in ways that are useful for some purposes, but they do not throw light on the hard
question, the actual nature of word meaning as it is computed by the mind and the
brain. In effect, a feature norm has characteristics of encyclopedic meaning and
transliterations of aspects of semantic gestalts, but as we have seen there is,
arguably, a deeper and more interesting level of word meaning, the configurational
structure that is formed from semantic factors, that is absent from this approach. In
addition SDFs are likely to make heavy weather of other types of words, qualic,
closed-class and abstract, in particular.

The compilers of the feature norms and those who have prepared similar lists of
concepts' features faced a formidable difficulty in that they were seeking to identify
constituents of word meaning with no prior valid empirical evidence of what their
nature might be. They have made what seemed a reasonable assumption – that the
meanings of concepts relating to living things and physical objects consist largely of
intuitively accessible features largely of a sensorimotor type. In the absence of other
evidence this must have seemed a safe assumption that is supported by anyone's
intuitive understanding of these concepts. These features seem to have impressive
psychological reality credentials.

Sensorimotor and allied information is undoubtedly implicated in the meanings
of concrete concepts, but the form in which it operates is open to question. There is
an alternative to SDFs, as we have seen. Sensorimotor information may be present
in word and concept meanings in the form of gestalts as I have described them in
Section 8. These are complex multi-modal cognitive entities with idiosyncratic
characteristics. They are holistic, they are ‘felt' mentally, they have an isomorphic
relation with what they denote and they possess distinct, identifiable, necessary
structural elements, the semantic factors. They are supplemented by idiosyncratic
sensorimotor material and encyclopedic meaning.

I propose that this approach has two distinct advantages over surfacedescriptive
types of features. The holistic character of gestalts means that they can
be cognized instantly. On the other hand SDFs are discursive and are not accessible
in the real-time of word recognition. Secondly, the semantic factors have an
independent proof of existence – their identification by articulatory iconicity.

Perhaps the chief advantage of the factor-structured gestalt model is that it
operates well with words of all grammatical forms, concrete and abstract, core and
near-core, although less well with specialist, less commonly used words as has been
discussed. This scope of application of the 3-part model is an enormous gain.

The supposedly empirical basis of SDFs is highly questionable. It has been
observed that when we attempt to look into the meanings of words we find what we
look for. The production of feature norms appears to be a classic case of this. There
is no question that the tested subjects recognized these features but it cannot be
concluded that they were, therefore, the basis of their knowledge of the words'
meanings. In line with the 3-part model I suggest that the features that are identified
in feature norm processes are either fragments of encyclopedic meaning or facets of
gestalts. The factoral structure of meaning goes unrecognized as it does with anyone
who is unfamiliar with the factors. In fact it would be very difficult to identify in
concrete words that are typically used in feature norm derivations, like ‘knife', ‘duck'
and ‘cucumber', because the meaningsof these Type 2 words are invested mainly in
their gestalts, with only limited factoral support, and in their encyclopedic meanings.
SDF investigations have, in effect, shone a light on the surface facets of gestalts, but
their value needs to be viewed in the sharper light of the 3-part model with its innate
dimension-like factors.

In sum, there have been many interesting developments in the theory of word
meaning since the 1960s but little agreement has been reached. Perceptions of the
nature of concept and word meaning have been deepened markedly by the work of
many researchers but in particular, I believe, by Fodor's evidently maverick theory;
Jackendoff's and Pustejovsky's grappling with the structure of word meaning and
co-compositionality; the highlighting of the embodiment of word meaning, the important
role of encyclopedic meaning, and image schemas in the Cognitive Semantics of
Lakoff and Johnson; and Barsalou's perceptual symbols. The feature norms of
McRea and his colleagues have some practical applications in neuropathology but
they cannot purport to provide a valid account of word meaning. A range of
alternative approaches thus lies on the table but both the architecture and content of
lexical meaning remains unclear in the absence of substantial agreement among the
authorities.

The only issues on which there is a measure of agreement are that word meaning
consists, at least in part, of semantic features of some type and that encyclopedic
meaning has an important role. What seems to be wanting is an explanation that can
deal with all types of words, that is psychologically and neurologically realistic, and
that has a sound empirical basis. None of the theorists discussed so far have met
these requirements to a significant degree.

We have seen that the semantic factors and the 3-part model of which they are
the core has provided a basis for telling criticism of recent and current approaches to
word meaning. I argue that this model provides a much more satisfactory explanation
than the existing theories on account of the empirical warrant that I have attributed to
the semantic factors that should be greatly strengthened by parallel investigations by
other linguists.


annotated contents




13. A new theory of word meaning.

Following on from the discussion in the previous section I will take the 3-part model
as the basis of a new theory of word meaning that has the power to explain major
issues regarding word and concept meaning. I believe the results exceed those of the
theories I have reviewed by a large margin. To this end it is useful to consider how
well it performs against a useful set of ‘desiderata' proposed by Prinz (2004), with
one amendment, the inclusion of ontology. What should a theory of word and concept
meaning attempt to achieve? The scope is large but it can be narrowed to seven
headings:

.......Ontology. How can we specify what the meaning of a word/concept actually is?
.......Scope. Can the theory account for the full range of types of word meaning?
.......Intentionality. What gives word meanings the capacity to be about worldly entities?
.......Cognitive functions. What functions do they serve in the cognitive and language economies?
.......Compositionality. What gives words their capacity to combine to form larger units?
.......Acquisition. How are word meanings acquired?
.......Publicity. How are they shared among individuals given that they exist independently in
.........individual minds?

In this discussion I am regarding the meanings of words and simple concepts (wordsized)
as roughly equivalent. One operates in language and the other in non-verbal
thought but they share the same three-part structure.

In the form of the model described in Section 9 we appear to have a basis for
addressing these matters. The key understanding we have gained of the semantic
factors is that they have a dimensional nature that represents isomorphically
dimensional aspects of the world as experienced by individuals. The factors' somatic,
body-related operation in the human mind/brain is the basis of the isomorphic
relationship. We physically sense/feel semantic facets of concepts as we physically
sense/feel real entities, but at a fraction of the intensity of the latter. Aspects of the
world that are crucial to human cognition and behavior can now be seen to operate at
the very heart of word meaning. We have unprecedented objective evidence of their
existence in their iconic articulatory representation in some languages and in the
confirmation that intuitive awareness of them can provide. The composite,
isomorphic, gestaltic element of word meaning with its factoral structure that
establishes words' core semantic value is both the organizing principle of the
idiosyncratic sense of each word and the intuitable ‘face' of word meaning.
Encyclopedic meaning, operating through the associative ‘innernet' of long-term
memory, provides a wealth of supplementary information that can be utilized in
specific contexts.

The lineaments of a theory are not difficult to isolate in this description. It is a
remarkably simple theory. Word meaning is not atomic as Fodor and others hold. It
has a constitution that can now be identified with confidence. The dimensional
constitution of lexical semantic structure is common to every word in every language.
It is an immensely powerful means of representing the world in the mind/brain, and of
denoting external entities. Semantic gestalts, on the other hand, have a distinctly
atomic appearance in our instantaneous awareness of them. But their atomicity is
transparent once we are aware of the semantic factors. The dual nature of the
semantic factors, alpha and beta, provides much of the power of words to create
meaning that is both graphically descriptive and endowed with the affectivity that is
crucial to the active combinatorial functioning of words, as it is to our perceptual and
cognitive experience. The beta factors extend the domain of word meaning into the
former semantic no-man's-land of abstract words. The factors also have a surprising
capacity to generate the grammatical words that have often been assumed to have
meaning only by virtue of their use.

Let us now consider the above seven issues so as to evaluate the performance
of the 3-part model. In doing so I do not have the space to discuss the controversies
that surround each of the issues. I will assume that the interested reader has some
familiarity with them and I will simply outline the response that is implicit in the model
as I have described it.

.......1...Scope. A satisfactory theory must be able to account for the full range of word
............meanings, simple and complex, closed class and open class, abstract and
............concrete. No existing theory has this scope but the current theory has the
............capacity to describe word meaning in detail in all these types. This emanates
............from the universality of the factors, the uniqueness of each semantic gestalt
............and the scope of each person's associative ‘innernet' of encyclopedic
............meaning. The theory can provide a full descriptive taxonomy of word meaning.

.......2...Ontology. How can we describe the ontology of word and concept meaning?
............This matter is complex on account of the existence of words in the heads of
............individuals, in the public domain of communication and presumably in an
............abstract global semantic space. The primacy of the dimensional factors and
............gestalts in word meaning arises from their specific source in somatotopic
............regions of the human brain, in neurons that serve tissues in parts of the body
............with specific functions (sensory, motor, proprioceptive and so on). It is
............arguably impossible to avoid the conclusion that this is the ultimate source of
............the semanticity of words, their capacity to bear information. The substance of
............word meaning, its semantic content, is contingent on this neural source with
............encyclopedic meaning contributing supplementary but necessary material.

.......3...Intentionality is a technical term that denotes the capacity of words and
............concepts in the mind and in use to be about something external to the mind. It
............was described by the originator of the concept (Brentano) as the unique ‘mark
............of the mental'. It has been viewed as a puzzling capacity and its validity is
............denied by a number of authorities. Given the ontology just described and the
............isomorphic nature of the factors and gestalts the puzzle dissolves.
............Intentionality is derived from the isomorphic relation between meaning as
............neurally inscribed in somatotopic regions of the brain and features of the
............actual world as it is experienced. It is important to recognize that intentionality
............is fundamentally biological as it applies to the capacity of any organism to
............register external entities and events in proto-concepts (that are bound to
............immediate sensory cues). It is not the unique mark of word meaning (or the
............mental). That lies elsewhere – in part in the unique human capacity to
............combine and manipulate concepts and words in the absence of immediate
............sensory cues.

.......4...Semantic functions. Word/concept meaning serves several semantic/cognitive
............functions. The most commonly recognized are denotation, reference and
............representation: the psychological capacity to denote external entities in a
............general sense through categorization; the capacity to refer to external entities
............in a specific sense in a language context thus contributing to the meanings of
............sentences; and the formation of mental symbols. All are contingent on
............intentionality as the source of semanticity and ultimately on neurological
............isomorphism. Some authorities (particularly Chomskyans) deny the relevance
............of these functions. They assert that semanticity should be regarded in a
............science of language as a purely mind-internal affair that serves the language
............user's capacity to denote and refer. This marks a deep divide
............(internalism/externalism) in contemporary linguistic thought. The 3-part theory
............indicates that internalism is not a valid approach because semanticity is
............primally and inalienably intentional. The external relations of word meaning are
............integral to their viability (see 15.3.8).

.......5...Compositionality. Language has the foundational capacity to combine
............words/concepts in an unbounded number of ways by utilizing the meanings of
............individual words and specific ‘rules', some intrinsic to the language faculty and
............some specific to individual languages. Compositionality underlies the
............boundless productivity of thought and language and its systematicity – the
............existence of inherent structures of combination that are common to thought
............and language. The model I have described advances two important
............considerations. First, compositionality is not confined to language and thought.
............It is an integral element of everyday experience involving direct sensory
............inputs. We do not perceive entities individually but in relation to other
............associated entities both in the moment and in the flow of time where causal
............relations often become evident. This interaction with our surroundings is
............intrinsically compositional, productive and systematic. The compositionality of
............language and thought is derived from this primal form in experience in which
............the factoral dimensions serve to integrate the components of experience within
............a meaningful composite structure. Second, in language and thought entities
............etc. are abstracted from the direct experiential composite flow and are
............individuated as concepts on the basis of their gestalts and their dimensional
............structures that operate more covertly in direct experience. Here the internal
............dimensional factors of words/concepts and facets of gestalts operate as
............valencies that establish combinatorial possibilities within phrases and
............sentences. This same process operates in both language and thought (LOT).

.......6...Acquisition. The theory needs to be able to give a plausible account of both
............how word/concept meanings emerged originally in the human mind and how
............they are acquired ontogenetically, by individuals, assuming as I do, that they
............are not innate and subject to triggering by experience. In terms of the new
............theory they are learned with aid from the innate alphabet of word meaning.
............Both issues are discussed in more detail in 15.1 and 2. In brief, human
............cognition is founded on the alphabet of meaning existing prior to discursive
............thought and language and its use in conceptualization prior to language. The
............unique attribute underlying language and discursive thought is, as we have
............seen, the ability to hold concepts independent of immediate sensory stimuli,
............and to combine them into larger concepts. This was the most crucial
............prerequisite for the original emergence of language and is for its ontogenetic
............development in every young child's language.

.......7...Publicity. Words and concepts are framed in individual minds on the basis of
............individuals' unique learning and experience. Meanings are far from uniform but
............satisfactory, often flawless, communication is the norm. The factoral core of
............words is common to all members in a language community. Semantic gestalts
............are also largely common to members. Encyclopedic meaning diverges
............substantially from person to person. Significant commonality is an advantage
............in communication, but the other two elements ensure sufficient semantic
............common ground exists to enable communication to occur satisfactorily most of
............the time.

Assuming agreement on the vital ontological question this 3-part theory of word and
concept meaning performs extremely well in terms of the desiderata. There is not
scope here for comparisons in this respect with the theories described earlier but the
interested reader may wish to consider the conclusions Prinz reached in his 2002.

The main features that distinguish this new theory from the other contenders are
the presence of the semantic factors and the postulation of semantic gestalts. I
expect the validity of the semantic factors will be accepted by linguists in due course.
The concept of semantic gestalts should also have considerable appeal on account
of their multi-modality that gives them a big advantage over images, image schemas
and perceptual symbols. With these two key elements plus the theory's unparalleled
performance in the analysis of word meaning the new theory would seem likely to
eventually gain support from linguistic authorities.


annotated contents




14. Neuropsychological research into word meaning.

14.1 Some recent investigations.
Over at least the last three decades or so there has been extensive research into the
neural correlates of word meaning, into how the contents of word meaning, whatever
they are, are represented in regions of the brain. This work relates directly to the
hypotheses that I am advancing in two ways: it has the potential to assist in the
identification of the neural correlates, the ultimate concrete evidence, of the semantic
factors, if it is to be found, and, on the other hand, the 3-part model seems wellplaced
to aid neuroscientists in interpreting data that they obtain in brain mapping.
This data is generally very difficult to interpret confidently with the consequence that
there are many conflicting conclusions.

Neuroscience research institutes, mainly in universities in the USA, Great
Britain and Europe are engaged in a diverse range of studies of language along with
other matters, especially relating to brain pathology. A voluminous number of
research papers is being produced each year on the neurological basis of language,
many using evidence provided by pathological or accidental neural deficits and
others using a range of means to study word representation in healthy subjects. This
field involves several complementary areas of work including the investigation of the
effects of impairments to brain function in specific regions, and the use of several
brain-scanning technologies such as fMRI (functional magnetic resonance imaging)
EEG (electro-encephalography) and MEG (magneto-encephalography) to study the
operation of the brain directly in cognitive functions. A considerable number of
studies that use these approaches focus specifically on word meaning. Numerous
groups are producing research that is endeavouring to draw some firm conclusions
on how the human brain constructs cognition and linguistic meaning – how the
knowledge of the meaning of a word is organized in the brain, how it represents the
meaning of laugh or persuade?

The use of these technologies for such purposes is relatively new. They only
became available late last century and fMRI has been used in language research
only over the last fifteen years. Both the equipment and the means of interpreting the
results are constantly being improved but the work is complex and difficult. The
technologies still have marked limitations but some promising results are being
generated. This is a challenging field that is advancing understanding of the
operation of the brain including its engagement in word and wider language meaning,
but a clear picture has yet to emerge.

It is impossible to review this work in any detail here because it is so extensive.
I will instead focus on one recent ‘meta-analysis' of a large body of work and more
detailed comment on one specific line of investigation that has something in common
with the approach of this paper. In a 2009 paper entitled Where is the semantic
system?
Jeffrey Binder and associates reviewed and analyzed 120 functional
neuroimaging studies of word meaning carried out from 1980 to 2007 out of a total of
over 500 published studies. This is one of a number of such meta-analyses produced
in the last few years with a view to clarifying areas of agreement arising from a range
of compatible studies. The fact that such meta-analyses are being undertaken is
indicative of the level of interest and activity in this work and its difficulty in reaching
agreed results.

The paper produced a broad picture of a gradual convergence of this research
in terms of identifying the regions of the brain that engage in ‘semantic processing',
the ‘cognitive act of accessing stored knowledge about the world' that is involved in
the understanding of word meaning. It found significant agreement on distinct
semantic sub-systems in the brain and their functions in relation to specific types of
words. Interestingly the analysis showed an almost full overlap between the semantic
system and the autobiographical memory system that contains knowledge about an
individual's past. This is attributed to the fact that the autobiographical system
contains stored knowledge that, ‘over time loses perceptual detail and comes to
resemble semantic knowledge'. It would be surprising if the two types of knowledge
were duplicated in the brain. But it is clear from this paper and other recent literature
that much remains to be done to gain a satisfactory understanding of the operations
of the neural mosaic involved in semantic processing. In concluding the paper the
authors observe that ‘future meta-analyses of cognitive imaging studies would benefit
from the establishment of a more formal ontological system for defining the cognitive
processes represented by ‘activation'. ‘The object….is to identify the neural
correlates of a specific cognitive process or related sets of processes, but this
enterprise cannot logically succeed without an objective means of defining the
cognitive components represented by an activation'
. (My emphasis). This seems
to point directly to what the semantic factors and the 3-part model have to offer in the
way of ‘cognitive components'. Binder's view is supported by Poeppel and Embock
(2005) who emphasise the need for ‘an integrated theoretical and experimental
perspective' if real progress is to be made.

This meta-analysis and many other papers focus largely on the where, that is,
determining the regions of the brain that are activated by various kinds of word
meaning. Other researchers are approaching the issue by seeking to understand the
how, the actual process of re-cognizing word meanings. I have selected the work of
two research groups with similar approaches for discussion. Their work has close
parallels with the subject of this paper as both envisage semantic features as forming
the structure of the how. How and where are intimately connected because
constituents of the ‘how' may be indicative of location.

The first group involves H. Moss, L. Tyler, L. Taylor and colleagues of
Cambridge University, Great Britain and the second, G. Vigliocco, D. Vinson and
associates of the University of London. Both groups utilize semantic features and
decompositional analysis in their investigations. This kind of analysis has continued
to be regarded as a valuable form of investigation of word meaning. It is being widely
used both in the investigation of the neural instantiation of word and concept meaning
and in attempting to better understand language deficits (agnosias) caused by
various types of damage to the brain. Its use has been fostered by its relative
success in dealing with words with concrete meanings, particularly nouns.

These two groups of researchers both use McRea et al's surface-descriptive
features (SDFs) described earlier. They have been characterized by Vigliocco and
Vinson (2007) as ‘not abstract, but grounded in perception and action, at least to
some extent. In other words, conceptual features, the building blocks of semantic
representation, are embodied in our concrete interactions with our environment.' It is
obvious that this puts a much more positive gloss on SDFs than is warranted when
they are compared with the semantic factors with their impressive range of
characteristics as described in 10.1. SDF norms have proved useful in some
neuropathological research but their scope seems rather limited. On the other hand,
the semantic factors and the 3-part model appear to have much greater explanatory
power and the potential for wide application in neurophysiological research into word
meaning.

14.2. The semantic factors as a platform for neuroscience investigations of
word meaning
.

The nature of word meaning is still a puzzle for linguists. Section 12 is indicative of
the breadth of views that are still current. There is an expectation that neuroscience
should be able to provide some helpful evidence of how word meaning is constituted
in the brain. However it is very unfortunate that linguistics has been unable to provide
a firm platform for use in neuroscience investigations. In particular there is still
uncertainty among linguists about the nature of the semantic features that are widely
regarded as key constituents of word meaning. The surface-descriptive features that
are often used have a ‘folk psychology' quality about them. While they are based on
commonsense intuition they seem to hold limited potential for generating insightful
results from neurophysiological research let alone for constructing an adequate
theory of word meaning. Investigators who work with surface-descriptive features are
seeking to investigate concepts largely via perception and the information it provides.
This approach ignores an important aspect of the world as we know it (the affective
factors), the component that actually generates meaning as opposed to mere
description. The results of the work summarized in this paper provide an alternative
that would appear to be worth testing despite its current lack of authoritative support.
The critical element that they contribute but that is poorly represented in norm-based
features is the affective beta factors that have an evaluative, behavior-related
function. A more satisfactory result, it seems, should be available by approaching
concepts from the cognitive route offered by the 3-part model, with the unique
opportunity provided by the semantic/cognitive factors and gestalts to investigate
meaning in a differently targeted manner.

It is time to explore in a preliminary way the idea that the 3-part model and
particularly the semantic factors may be able to assist in understanding the operation
of the human brain in the formation of word meaning. My perspective here is
linguistic but I am seeking to engage with neuroscience. Researchers in both fields
acknowledge that a joint approach is likely to be the most productive. My own work
has led to the same conclusion. I have drawn attention from the beginning to the
factors' apparent capacity to form a bridge between our knowledge of word meaning
and the neural mechanisms that generate it. It is unique among theories of word
meaning in this respect. If this conclusion is valid it is a significant advance. The 3-
part model has a number of features that distinguish it from other approaches and
that seem to have substantial potential to be used by neuroscience researchers in
productive ways.

First, it has the distinct advantage that its components, particularly the factors,
can be identified in language in an empirically robust manner, with articulatory
iconicity as the vital clue, but once they have been identified in this way their much
more common non-iconic presence can be identified in any word. They operate
effectively across a far greater range of word meaning types than the features that
are the mainstay of current neuroscience investigations.

Second, as we have seen, the model can deal with any type of word. The
investigation of different types of words, for example those with the range of affective
contents represented by the beta factors, may be enlightening.

Finally, and most significantly, with their dimensional nature, the factors seem
to be active in fusing a linkage between denoted objects and events and other
representations with specific regions of the brain. I have described the semantic
factors' representation of somatotopic aspects of word meaning – sensory, motor,
kinaesthetic, proprioceptive and interoceptive or affective in 10.2. There is evidence
that every one of the 24 factors has a credible association with specific somatotopic
areas of neuronal activity. This systematic somatotopic structure of the factors in
relation to word and concept meaning and other forms of cognition would appear to
have great potential to provide the kind of guidance that Binder et al have found
wanting. On the face of it at least, they appear to provide vital clues to how word
meanings are represented in the brain. This appearance may be deceptive, but only
investigations into the factors evident role can decide the matter.

As I have stressed on several occasions progress in this direction depends on
two further developments. First, the material set out in this paper needs to be peerreviewed.
Second, some initial neurophysiological investigation of the operation of
the semantic factors needs to be conducted. If this produces positive results further
work will hopefully follow.

There is however reason to be cautious about the likely rate of progress. I
have mentioned the difficulties that neuroscience labours under. This is reinforced by
two recent books (Satel and Littlefield (2013) and Barton (2013)) that highlight the
inherent problems associated with the current technology and cast serious doubt on
the validity of much recent research. It is to be hoped that these issues can
eventually be resolved.


annotated contents




15. Some implications of the semantic factors for linguistics.

This papers' discussion of the function of the semantic factors in language has been
entirely about their role within the meanings of words. They provide an entirely new
way of understanding the constitution of word meaning. It is very unlikely that such
basic elements of language are unimportant beyond the boundaries of lexical
semantics. In this section I will discuss some wider aspects of language in which I
believe they are profoundly implicated: the evolutionary emergence of language, the
learning of language by young children, the nature of syntax and grammar and the
factors' possible implication in another foundational aspect of language, the
combination of words in phrases and sentences.

A large caveat is necessary at this point. These are fields in which there has
been massive scholarly investment and achievement over the last half-century or so
but they remain highly controversial. I cannot claim to have absorbed more than a
fraction of this work although I believe I understand much of its central drift. In these
circumstances my attempt to contribute to these fields may seem foolhardy, to say
the least. My attempt is motivated by my belief that the discovery of the semantic
factors creates a unique opportunity for new insights to be gained in these areas.

I have described the semantic factors as the alphabet of word meaning. There
are two other metaphors that capture something of the scale ans scope of this
opportunity. We can view the factors, adapting a phrase from information technology,
as the source code of the mental/neural computation involved in the generation of
the word meanings on which the construction of phrases and sentences is based. As
a source code the factoral alphabet is a ‘fully executable description of the software
system' that operates the neural hardware that is responsible not only for language,
but also for cognition and behavior. This code is written in the ultimate language of
biology – the sensori-motor parameters and dispositional principles of somatic
interaction with the environment that I have described. As a variation on this theme
we can regard the factors as the DNA of language. It is, I propose, the blueprint for
much of the everyday operation as well as for the emergence of language in both
evolutionary and child development terms. In the evolutionary time-frame the
blueprint was always there, operating at the interactional level of organisms. Its
language function was triggered once a range of other necessary physiological,
neural and cognitive functions were present in Homo sapiens. Children are born with
the factoral capacities and utilize them as they learn their first words and
subsequently as they grow their lexicons. Equally importantly, I believe, the factoral
DNA is indispensible for the process of the construction of sentences and the
discursive uses of language (15.4 and 15.5).

These metaphors highlight the evident level of significance that these entities
that have lain unrecognized in the lexicon appear to be capable of bearing.

15.1. The emergence of language.
Theories of the emergence of human language continue to proliferate but convincing
explanations remains elusive. It is a hard problem, and a ‘hardy annual'. Many
linguists, including this one, cannot resist theorizing on the origin of language. Can
knowledge of the semantic factors contribute something useful?

It is useful to distinguish several probable phases in the emergence of language to
account for this obscure but vitally important transition in evolutionary development.

........1. Prior conditions. Early pre-linguistic hominids, like present-day animals,
............were capable only of immediate perception/cognition of experiential tokens
............(one-off sensory images) that were available only in the presence of direct
............sensory perception. These were in the form of sensory stimuli operating on the
............framework of the alpha factors plus the beta factors that were/are necessary
............to motivate response. The memory that is prominent in many species is simply
............recalled tokens that are responded to but not manipulated.

........2. Transitional conditions. The critical development was the decouplement of
............cognition from immediate stimuli enabling the cognition of types (abstract
............generalized categorical concepts) in the form of coded information – factor
............configurations and gestalts. The transition from token cognition to type
............cognition would seem to have involved a ‘stripping down' of neurally expensive
............tokens into much more abstract and more neurally efficient type-concepts,
............stored in long-term memory, that allowed generalization across many tokens.
............This process required the reduction of sensory content to a bare essential
............minimum and a consequent elevation in the prominence of the cognitive
............factors associated with a concept. A contemporaneous expansion of the
............memory for encyclopedic meaning would have been required. The timing of
............these developments is very uncertain.

........3. Discursive non-linguistic thought. Type concepts could, in due course, be
............combined in individual minds to construct more complex concepts, thoughts,
............the combination being effected largely by the valencies (largely associated
............with the now more salient semantic factors) inherent in individual types to
............produce more complex mental artifacts. This development would probably
............have had a marked adaptive benefit through an enhanced capacity for
............interaction with the surrounding environment. Such combined concepts would
............have been a form of Proto-Language of Thought (LOT) (see below) in which
............individual concepts had the 3-part model form.

........4. Emergence of proto-language. At this phase neurally-encoded type-concepts
............became associated with signs (spoken words or manual and bodily signs). A
............natural form of association was via iconic kinds of signs that could have
............facilitated communication initially. Increasing arbitrariness must have occurred
............as vocabularies grew and the scope for iconicity diminished. The initial protolanguage
............may, however, is likely to have remained heavily iconic if the
............remnant iconicity in today's languages is anything to go by.

........5. Emergence of discursive language. This was, I believe, a less demanding
............phase. But what was the source of the syntax of sentences? I propose that it
............was derived from the natural syntax of the LOT where it resided then, as now,
............in large part in each word's factoral/gestaltic template that helps to select the
............words that can be associated with it. As communicative demands and abilities
............developed new words and syntactic artifacts would have emerged to perform
............necessary grammatical functions (e.g. closed class words) and prototypical
............combinatory patterns would have developed. These could be learned to
............facilitate sentence construction to meet standard needs (see 15.2 and 15.5).
............Language as we know it had emerged.

The second phase, the emergence of manipulable type-concepts was the most
critical in the whole process and may have been very protracted. It would eventually
have produced a new dimension of cognition, the ability to think of entities that were
not immediately present – the ability to think as described in phase 3. Such thought
would seem to have been in the form of a Proto-Language of Thought. (I use Fodor's
term for both this Proto-LOT and also the LOT that is the medium of our own preverbal
and non-verbal thinking. I use it without his extensive philosophical underpinnings.
In my usage the LOT is the format in which the brain computes the
meanings that are expressed innon-linguistic, intuitive thought, linguistic thought and
language. It is an intermediary between the brain and language, an intermediary in
which the factors exercise their alphabetical function in the mental representation of
concepts (and eventually in word meanings). This is a far cry from Fodor's heavily
philosophical approach, but it shares his emphasis on the representational nature of
concept meaning and the computational nature of thought). I conceive the Proto-LOT
as a fundamentally neural form of thought in which the semantic factors and gestalts
operate without crystallization into words. The Proto-LOT is the language of gists
(highly schematic forms of meaning) rather than specific words. But these gists
contain the templates of the verbal sentences that may eventually emerge from them.
The LOT is written in the neural language that occurs in somatotopic regions of the
brain (10.2). For us, gists can be intuited in a felt sense but they are normally swiftly
transformed into verbal thought or language. The Proto-LOT as I conceive it was a
purely conceptual, non-linguistic form of thinking, detached from immediate cues, and
possessing a propositional character to some degree, that preceded the emergence
of language (as LOT formulations arguably precedes our linguistic thought and
speech). Evidence of the existence of the Proto-LOT lies in the significant
development of technological and social practices of later pre-linguistic hominid
species that archaeology has revealed.

These achievements presuppose a propositional thinking capacity, respective
and prospective
, drawing on experience to plan and coordinate these practices. The
semantic factors were the structural alphabet of the Proto-LOT concepts and
propositional schemas. Once this was in place, it would seem natural that increased
communication between individuals would have followed but it is not possible to know
with much precision what neural or cognitive limitations needed to be overcome for
this to happen, and what timeframe this process would have followed.

This fourth phase, the emergence of Proto-language, is normally regarded as
the key to the emergence of language but from my perspective it was a less
significant evolutionary development than the emergence of type-based
conceptualization that is arguably the most critical distinction between human beings
and other animals. Early communication would have been in the form of a refined
and enlarged gestural repertoire or a vocal proto-language, or a blend of both.

What communication required was sound/gesture structures that represented
LOT concepts in ways that could be understood by conspecifics. The semantic
elements of combinable concepts – factors, gestalts and encyclopedic meaning –
would have already been present in Proto-LOT symbols. A very natural means of
externalizing them, converting them to a communicable form, was available. This was
iconicity, the utilization of a similarity of concept meaning, or at least a signature
aspect of it, on one hand, and analogical gestures or sounds on the other.
Articulatory or gestural iconicity is likely to have been much more pervasive in a
proto-language than in some modern languages and would seem to have been ideal
for this purpose. It would have been equally potent for the maker and receiver of the
signs, both LOT users. The principle of iconicity was not new. It would have occurred,
as it does now in vocal expressions and gestures that were common in the animal
world. Authorities cite low growls as conveying the sense of largeness in aggressive
animals and cringing body shapes as expressions of submission, but there are many
other instances. The use of iconicity in a much more complex way to create vocal or
gestural symbols with the semantic content of LOT ‘words' would seem to have been
a natural development, a bridge from LOT to language. Even at this stage, however,
words could not have been fully iconic, although gestures, by their nature, could have
been more so. The principle of Saussurean arbitrariness would have been a
constraint on iconicity. Partial iconicity could, nonetheless, have been a key catalyst
for the emergence of communicable words.

I believe this is a feasible scenario regarding the emergence of a very basic
word-based proto-language without significant syntax, that would have had
considerable practical utility. This would have been an enormous step in the evolution
of language, but severely limited in terms of its usefulness without the means to allow
more complex phrases and sentences to be constructed. I will not speculate on this
phase here (see 15.4) beyond noting that the emergence of words of the main
grammatical categories, nouns, verbs and adjectives would have created powerful
valencies that would have facilitated the fundamental processes of predication and
combination and the formation of phrases in word orders that would have become
fixed by a combination of convention and logical and pragmatic considerations. This
would seem to describe a viable language emergence process with iconicity as a
valuable catalyst in achieving intelligible communication.

15.2 Language learning by young children.
I discuss this contentious topic, also, mainly on the basis of the material I have
developed in this paper and without specific references to the voluminous literature.
This will be another test of the validity and robustness of the matters I have outlined.

A consequence of the existence of the dimensional semantic factors is that
they are involved in the computational processes of the infant mind from birth, the
processes that will enable the infant to generate the concepts of its proto-Language
of Thought well before the first words occur. Children are, in effect, born with the
dimensional source code of language at their brains' disposal. The semantic factors,
in a cognitive form, are necessary for a child to make sense of its experiences with
the alpha and beta factors combining in giving meaning to percepts and protoconcepts.
Prelinguistic childrens' brains are able to compute the meanings of
emergent concepts – mother, face, breast, eyes, hand, milk, blanket, toy, shoe and
so on – in an infant proto-LOT. These are formed, like our own, as gestalts built on
the factors' dimensional frameworks and associated with the bare beginning of
encyclopedic meaning. (Even very young children begin to develop an ‘innernet' of
resources that facilitate the use and combination of concepts and words). It must take
a while for the child to form the categorial concept shoe or toy from its experiences of
various examples, but it gradually forms a conceptual lexicon well before it begins to
assemble its verbal lexicon. The former is arguably a necessary resource for the
eventual development of the latter.

The word-learning of young children has several other things working in its
support. First, young children have the great advantage over the first human
speakers, that they are exposed to language from the beginning of their lives
(perhaps before to some degree). They live in an envelope of language from birth.
This enables their brains to become attuned to the phonological repertoire of the
surrounding language very early and to the idea that small clusters of sounds are
significant in some way. As early maturity proceeds children seek to express salient
LOT elements, frequently with the assistance of parental ostentation and repetition.
The child must become increasingly aware of spoken words that often co-occur with
what they represent in ambient and child-directed talk. This may be reinforced
innumerable times in these critical months. Second, as with the first human speakers
or signers, the development of the infant LOT of concepts with gestaltic forms and
factoral structures provides the semantic basis for the early words. Learning ‘drink',
‘cup', ‘shoe', ‘on' and so on requires an appropriate configuration of structural factors
and also an associated cluster of sounds that caregivers are generally keen to
provide. Once the trick is learned words begin to multiply. Third, the presence of LOT
concepts (symbols without sounds) in the pre-linguistic infant mind may operate as a
spur or motivation to the expression of these concepts that it has been observing
being expressed by adults. LOT concepts and ambient speech feed an expressive
urge. Fourth, children have an ability to, not only imitate adult language behavior, but
to mime it, to replicate it in knowledge of what it means and with the intention of
conveying that meaning and obtaining the reward of recognition.

With these pillars of support for the development of language in young infants
the process appears eminently natural and unmysterious. I do not believe articulatory
iconicity to be such a major factor in the development of the child's vocabulary, as I
have speculated that it did for the first hominid language users. Iconicity is less
prominent in the early words a child learns than I proposed it would have been in the
newly-coined words of the first proto-languages. However it provides significant
1support in the ongoing acquisition of the core lexicon in which it is probably pervasive
in most languages.

The above elements would seem to go a long way towards accounting for the
learning of single words and mundane canonical phrases, but how does a child learn
the elements of syntax and grammar, the morphological forms of words, agreement,
tenses and so on and the appropriate formation of more complex structured
sentences? Do we need to postulate an innate Universal Grammar (UG) with innate
principles as Chomsky and the many adherents to his theories have done?

By the time the child begins to form sentences of gradually increasing length it
has been continually exposed to the distinctive patterns of phrasal constructions of its
mother tongue in the envelope of language in which it lives. It has been picking up
hints of their construction from isolated words it recognizes. A particularly potent
aspect of the development of syntax is the presence of natural valencies in words,
particularly verbs, that facilitates the linking of appropriate words in short sentences.
For example, the argument structure of verbs and the intrinsic grammatical functions
of all words are integral features of their natural meaning that are intuitively
accessible to a child. These have a strong influence on how words are assembled in
sentences. The factoral structure of word meanings also provides important clues
that facilitate the combination of words. In addition, the language's common wordorder
patterns that a child intuitively comes to learn, provide a repertoire of multi-word
constructions that it can learn with minimal difficulty. For example in learning to say
‘[Mummy put] shoe on [foot]' as its first sentence, the child intuitively knows that
putting involves three intrinsic arguments, an actor, an object and a position, that
determine the structure of the sentence. The argument structure of verbs soon
becomes a natural part of a child's combinatorial practices.

There is another influence that is equally significant. Children are constantly
exposed, throughout their development, to patterns of word use with their built-in
syntactic structure. This, in conjunction with the valencies of individual words,
provides the emerging complex architecture of sentences. As they continue to mature
children and adolescents become capable of understanding intuitively and mastering
the more complex types of syntactic structure that accounts of generative grammar
describe copiously.

In these circumstances there is arguably no significant ‘poverty of stimulus'
such as Chomsky and has colleagues have insistently postulated as being the preeminent
feature of the child's learning situation. The scenario I have described may
be sufficient for the child to master the grammatical, morphological and syntactic
patterns of its mother tongue in its speech and comprehension. The justification for
the Chomskyan innate Universal Grammar as operating in a child's learning of its
language begins to seem highly questionable. The power of the semantic factors and
gestalts in forming concept and word meaning, and the inbuilt valencies of concepts
and words, together with childrens' ability to learn key sentence constructions are
arguably adequate stimuli for the infant acquisition of complex language forms. This
approach will be anathema to generative grammarians because poverty of stimulus is
crucial to the existence of a substantial Universal Grammar.

15.3 Implications for Generative Grammar.
Theories of syntax and grammar have been dominated over the last half-century by
Noam Chomsky and the vast number of linguists, worldwide, who have, more or less,
subscribed to his views on language. The purpose of this section is to explore
tentatively the implications that the approach described in this paper has for
Generative Grammar and in particular for the Minimalist Program, its latest version.
This discussion is tentative because my familiarity with the voluminous Minimilist
literature is limited and the extent of the implications is hard to determine at this stage
except in broad terms. The purpose of these comments is simply to suggest some
apparently fundamental ways in which the approach outlined in this paper appears to
relate to the latest Chomskyan model. My justification, again, is that the semantic
factors, understood as the ‘source code' of language, appear to be profoundly at odds
with Chomsky's program in a number of ways – to the detriment of one or the other.

There are four basic premises to my approach here:

........1. A satisfactory knowledge of the nature of word meaning that is absent from
............Generative Grammar is fundamental to understanding the full operation of
............language from its neuronal source to its expression in sentences.

........2. Some elements of words that Generative Grammar often takes to be integral
............to an autonomous syntax, such as grammatical categories and the argument
............structure of verbs, are intrinsic elements of word meaning.

........3. All words have powerful inalienable intrinsic valencies that influence their
............combination with other words in sentences. These are partly expressed in
............semantic factor configurations.

........4. In the form of the semantic factors, a primal biological logic, in which the
............beta factors have a central role, is active at the core of word meaning. It has a
............powerful influence extending into the combination of words in phrases and
............sentences that is generally held to be in the province of syntax.

A central aspect of Chomsky's approach from the outset has been the subjugation of
the role of semantics and an elevation of the role of syntactical rules viewed as
operating autonomously in the formation of sentences. This has been modified
somewhat in later versions of Generative Grammar but at least one authority is
currently still proposing that sentence meaning is inherent in the syntax of
expressions (Hinzen 2007). The idiosyncratic meanings of words have not assumed
a significant position in Chomskyan thinking. Words are often viewed as ‘atoms'
manipulated by syntax. Word meaning has been left to philologists and
lexicographers while grammarians have sought to unveil the logical essence of the
syntax of sentences, presuming it to be the central, determinative, defining element
of language.

15.3.1 Rationalism, logic and language.
Rationalism, and especially that of Descartes, provides the philosophical basis of
Chomsky's approach to language. From his first book to his latest writings
mathematics and logic have been central to his way of thinking about language. The
logical structure of sentences has priority over the semantics of their constituents. An
innate logic is held to be the definitive element of language. The dimensional factors
suggest a different perspective: perhaps it is not the logic of logicians,
mathematicians and philosophers that underpins the structures of language, but
rather the ‘biological logic' of the basic human dispositions and behavioral
judgements that are fundamental to cognition, that are largely embodied in word
meaning and that, arguably, motivate linguistic expression and the ways words
combine. This is a ‘logic' beyond logic, perhaps. It is inherent in language as one of a
number of means by which the human organism seeks to advance its interests in the
world. I am thus not just posing an Empiricist principle against Chomskyan innate
logic, but something that is arguably more fundamental than grammatical principles –
the dimensional ‘source code' that governs biological behavior, including linguistic
behavior.

This is not to deny the importance of logic to language, but I am asserting that
this, too, with its expression in the logical form of sentences, has biological roots. We
can begin to perceive these in the factoral nature of the logical operators used in
formal logic and manifest in the words' use in language. These are evidently primal
operators in human cognition that arguably had a placed in the Proto-LOT of early
humans and were adopted by language in an early phase of its evolution. For
example key operators such as if, and, not, or, some, all can readily be expressed in
factoral configurations that reveal their biological origin and character. The nature of
the logic embedded in sentences is currently controversial (see, for example, Crain, 2012).

15.3.2 Syntax and semantics.
Both syntax and semantics are central to the operation of language but generative
grammar has long been noted for its ‘synactocentrism' and its down-playing of role of
semantics. In effect, Chomskyans' ignorance of the ‘source code' of the computation
of lexical meaning and the influence of the contents of word meaning on the
combination of words has created its undue emphasis on syntax. By contrast,
knowledge of the dimensional factors and the model of word meaning they imply
requires a diametrically opposite approach that views the dynamics of sentence
formation as being located to a large degree in the words of the lexicon with their
internal structures. Word meaning operates in conjunction with global linguistic
considerations such as case, agreement, pragmatics and mental efficiency and the
large number of idiosyncratic conventions and conventional constructions adopted by
each language. Key aspects of words that are often attributed to syntax, such as
grammatical categories and argument structure, are intrinsic to the meanings of
words and operate from that position. These plus the natural valencies of all words
are powerful influences on word combination in sentences. Lexical semantics and
syntax are intertwined.

15.3.3 Word meaning and sentence construction.
Chomsky is uncharacteristically tentative about the nature of word meaning. Perhaps
more than any other linguist, he has an intuition of the significance of the internal
constitution of word meaning in the form of semantic features, but he has commented
on several occasions on the extreme difficulty of identifying them beyond syntactic
features and a few others. He sometimes places it in the realm of the impossible.
This inscrutability has been prised wide open by the discovery of the semantic
factors. The Chomskyan ‘edge' features that are claimed to be important to word
combination are arguably identifiable with elements of word meaning that can now be
confidently specified. It is not the ‘edges', but the semantic core of words, that is a
dynamic force in word combination in sentences. Knowledge of this lexical structure
suggests another approach to word combination that gives words' idiosyncratic core
features (the factors) a key function. The implications of this for Chomsky's position
would seem to run deep (see 15.4 and 5).

The most basic syntactic operation in Chomsky's Minimalist Program is
Merge, the combination of two words or ‘syntactic units' to form a larger meaningful
entity under the influence of ‘edge features'. Successive merges form a sentence in a
bottom-up operation. Merge is purported to be the most basic generative process of
language. This single mechanism has become the centrepiece of syntax in the
Minimalist Program. It is the chief element of Chomsky's fined-down Universal
Grammar but its workings are still unclear. Knowledge of the semantic factors may
contribute to resolving the mechanism of combination and whether it is a bottom-up
process, as Merge implies, or of a top-down nature that the 3-part model seems to
suggest. Under this scenario a quite different approach to the construction of phrases
and sentences is suggested. In brief, it is a top-down process in which the
sentence/phrase unit occurs in the LOT first as a composite gestalt (a monadic gist)
that must be verbalized and linearized for expression in language (see 15.4 and 5).
Knowledge of the factors and the 3-part model thus raises major questions about the
role of Merge. The inner dynamics of words' meanings are postulated to be the key
influence on word combination in sentences.

15.3.4 Innateness, Universal Grammar and Universal Semantics.
The concept of an innate Universal Grammar, UG, has long been the definitional
feature of generative grammar together with ‘poverty of stimulus' that UG requires for
its viability. All linguists agree that some aspects of the language faculty or
supporting mental competencies are innate, but the question of which remains open.
As described here, the semantic factors, as the ‘source code' of the language faculty,
are innate in a specific, dimensional, sense. This, together with the proposal that key
drivers of word combination are inherent in words themselves, poses a large question
as to the need for an innate UG. When the semantic factors are recognized as a
virtual Universal Semantics operating across the human lexicon and motivating word
combination across all languages, the concept of a UG becomes dubious.

15.3.6 Poverty of stimulus.
As noted above, this notion has been a central supporting argument for the existence
of a Universal Grammar. If children receive from the environment seriously deficient
information on the grammatical construction of sentences such information must, it is
claimed, be innate. It is hard to conceive UG without poverty of stimulus. It has
become an essential element of Chomskyan ideology, a figurehead in any exposition
of Generative Grammar. In the light of knowledge of the factoral source code of
language and the role of learned conventional syntactic constructions, however, the
poverty of stimulus hypothesis looks vulnerable. If children's pre-linguistic LOT
contains concepts that have the range of valencies that words possess, the
construction of sentences utilizing the level of stimulus that is available from ambient
conversation and learning begins to look eminently viable. Without this plank UG
becomes even more open to question.

15.3.7 The biological foundations of language.
Chomsky frequently uses this phrase, and more recently biolinguistics, with reference
to the supposed genetically-determined UG, the innate component that he believes
makes language possible. It is, he claims, equivalent to a language organ operating
in the brain. He has argued vehemently against critics who believe that the late
genetic change that he proposes as underpinning UG is contrary to biologists'
understanding of evolutionary process. It is ironic that another key biological aspect
of language that is much more tractable empirically has, it seems, escaped his
attention, perhaps as a consequence of his monocular focus on grammar and syntax.
This is the small set of alphabetical factors that can be described as the DNA of
language and that is arguably the source of a previously unrecognized property of the
language faculty, its capacity to form a linkage between the somatotopic ‘language' of
the neurons and human languages.

15.3.8 Internalism and externalism.
An internalist approach seeks to explain aspects of language by identifying
appropriate mechanisms ‘inside the head' that account for the ‘external' operations of
language, such as it uses in representation, reference and denotation and in
communication and thought, rather than using these aspects as key explananda (as
they are used by Fodor for example). This latter approach is termed externalism.
Chomsky regards generative grammar as a strictly internalist enterprise. It does not
study language as it is commonly conceived as the vehicle of communication, but
i-language, the ‘individual, internal, intensional' (relating to meaning as opposed to
reference) aspects of language. The grammar's functions and its performance by
users are irrelevant to the understanding of how language operates. Only how
the mysterious internal grammar operates is of interest. Linguistics is viewed as a
form of psychology and at the same time as a natural science with key linkages to
biology. Its prime object is investigating how language is generated within the mind,
within the language organ. The external functions of language such as reference,
representation and communication are regarded as secondary – ways in which this
internal faculty is employed by language users.

For Chomsky the real science of language is the study of i-language and the
mehanics of the language organ. This is one of the most controversial aspects of
Generative Grammar. Many other linguists prioritize the external aspects of language
that are amenable to study in a great variety of ways compared with the relative
intransigence of language as it operates ‘in the head'. Chomsky's confidence in
tackling i-language stems in large measure from his dogmatic article-of-faith that an
innate Universal Grammar is the key to understanding language and that UG can
only be investigated ‘internally', as it is reflected in the form of sentences. UG, he
holds, is a consequence of late genetic changes in human beings that formed the
language organ and made language possible (another highly contentious matter). If
Chomsky is right in these respects an internalist approach may be valid.

What light can the 3-part theory throw on this matter? As the theory has the
innate semantic factors at its centre and as gestalts are psychological entities, it has
a distinct internalist appearance. However, the factors in their dimensional reality are
not intrinsically psychological. They are the necessary basis of the psychological.
The alpha factors have an external existence as physical parameters of the world.
The beta factors are biological behavioral parameters. Unlike UG, neither type of
factor is unique to our kind. Both are eons older than any UG. The central thrust of
this paper is that the factors are a bridge between brain and world that is essential to
all forms of human cognition and to language. The parameters of this bridge can be
identified empirically and are available for investigation by neuroscience. This is not a
language organ but an integral aspect of the operation of the brain. UG pales, it
seems, in comparison.

The theory of word meaning that I have developed cannot be designated as
either internalist or externalist. Is it complemented by UG at the other prime level at
which language operates, the construction of sentences in thought and in the
physical forms of language? I have argued that it probably is not. Other means of
combining words in sentences are available, not least the natural valencies that are
inherent in words themselves plus the multiplicity of learned conventional
constructions that arguably form much of the fine- and coarse-grained structure of
sentences that is normally attributed to syntax.

In 9.3 I have described the isomorphism that is inherent in gestalts and their
constituent dimensional factors. The somatotopic regions of the brain construct the
meanings of core and many near-core words (those with conceptually simple
meanings as opposed to the complex and often composite meanings of specialist
words). This conception of word meaning seems diametrically opposed to that in
Chomskyan thinking. These meanings have an intrinsic relationship with the external
world, a relationship that, I hold, is eminently biological and compatible with scientific
investigation.

The approach to language developed in this paper has a radically different
perspective from that of Generative Grammar and it seems to provide a potent
critique of the generative approach to language. The Universal Semantics of the
factors and gestalts provides at least a competitor for UG, and perhaps a wellfounded
alternative. In effect it offers a new approach to the science of language, a
re-engineered biolinguistics.

This critique of Generative Grammar is not intended to deny the value of the
unprecedented descriptions of the grammatical and syntactical machinery of
language that have been developed over recent decades by Chomsky and his
associates, but rather to question the validity of the basic explanations that posit the
UG as the ultimate foundation of the language faculty, on which it has always prided
itself. If my speculations are sustainable they seem to pose a significant challenge to
key aspects of the Chomskyan program but this challenge lies in limbo until my own
approach has been investigated and critiqued by my academic peers.

15.4 An alternative approach to the construction of sentences.
Taking the preceding discussion into account, this sub-section is a tentative proposal
regarding the means by which the human mind constructs sentences. It is intended
as an exploratory alternative to the Chomskyan hypothesis with the Merge function
operating from the bottom up.

The construction of words by combining sounds and elements of meaning and
the construction of sentences from words are absolutely central aspects of language.
It would seem unlikely that they are wholly independent and autonomous. I hold that
they are tightly integrated.

It is useful to begin at the level of thought, whether linguistic or alinguistic,
where there are parallel operations of the formation of concepts and their
combination to produce complex thoughts. Linguistics and psychology are at a loss to
explain satisfactorily the generation of simple concepts and the meanings of words.
The formation of sentences is usually taken to be a function of grammatical and
syntactic rules. The successive models of generative grammar have sought to
describe this function through a variety of mechanisms with Merge now in a prime
position.

I want to explore an alternative approach in which the construction of
sentences is largely driven by features of words themselves in association with a
range of non-lexical elements such as (i) the crystallization of conventional word
orders and phrase and sentence constructions that arise as languages evolve and
that are learnt by children and adolescents; (ii) the constraint of the necessity to
optimize neural efficiency including memory and computational limitations on the
complexity and length of sentences; (iii) ‘naturalness' and effectiveness of linear
organization that sometimes takes iconic forms (Givon 2011); (iv) the constraint that
common sense imposes in preventing the formation of nonsense sentences; (v) the
natural logic that is evident in everyday human experience of the operation of the
natural and social world; and (vi) the necessity for case and other types of agreement
to enable relations within the denoted world and the speaker's intentions to be
represented.

A key aspect of word combination is the initiatory or motivational stage that is
presumably some form of neural ‘pre-planning' in non-linguistic cognitive terms of
what a phrase or sentence is to contain. There must be some form of initial neural
expressive impulse with a broad semantic content. This applies equally to thinking. A
chain of thoughts also requires an initial substantive impulse. It may be stimulated by
the previous thought but it must contain new content if the train of thought is to
advance. The impulse may be in the form of a need for the natural (bio)-logical
progression of the thinking or the provision of a resolution. Chomsky has proposed
that in language this impulse is a belief, but there is a need to understand its content
and mode of existence more precisely.

Neurophysiological techniques such as fMRI have had some success in
illuminating the neural operations involved in the meanings of individual words but the
initiation and construction of sentences has so far proved impossible to investigate
neurophysiologically. The neurophysiological literature on the construction of
sentences is very meager in comparison with that dealing with word meaning. The
reason is that the initiatory and construction processes are effectively buried in a
mass of other neural activity. There are formidable methodological difficulties in
conducting the necessary investigations. In addition there has been little theoretical
work on this important stage.

One of the few attempts to model this process is outlined in Vigliocco, Tranel
and Druks (2012, Section 3.1) using a model in which the initiatory process is
described as message representation that is conceived as a ‘non-linguistic process
(e.g. information about the visual environment, encyclopedic knowledge, the
discourse record and a person's intentions)…that converge in preparation for verbal
expression'. This level is described as ‘impoverished' in representational terms, but
somehow attuned to the speaker's language in order to guide later lexical retrieval.
At the next stage, the functional level, words are retrieved and formed into sentence
‘frames'. This is a lexical and syntactical level. The paper specifies a clear role for
‘lexico-semantic representations‘ alongside syntax. Representations at this level
‘honor syntactic relationships among words'. A final positional level is ‘concerned with
the mapping between a hierarchically specified representation to a linearly ordered
frame that is phonologically spelled out'.

This model looks promising in broad outline but it is interesting to consider
what knowledge of the 3-part model might contribute. Importantly, it allows us to be a
lot more specific about the nature of the ‘impoverished' initial message. This is the
level where neuronal activity has a non-linguistic cognitive form. We can now
conceive it as having factoral and gestaltic elements, but of a non-lexical kind. I will
(tongue in cheek) take as an example Chomsky's well-known, curious illustrative
sentence from his seminal 1965 book: ‘Sincerity may frighten the boy'. We have no
idea why he chose this sentence, but let us suppose it arose more or less
spontaneously as a result of some unconscious stimulus (a belief?). In this case a
message emerged at the neuronal level of his brain with the gist of the sentence's
content. I suggest, to be a little more specific than just a message, that it was in the
form of what Johnson-Laird (1983) described as a ‘mental model'. These are, in his
scheme, ubiquitous cognitive constructs that are isomorphic with the state of affairs
they represent, and are constructed from innate conceptual primitives and semantic
operators
that represent relations between entities (time, space, possibility, causation
and others). Mental models are embedded in propositional attitudes such as belief
and desire. I propose that messages have this form.

The factoral and gestaltic content of the mental model of Chomsky's sentence
would thus arguably have consisted of a belief-motivated coherently structured gist or
template of its meaning made up from sufficient semantic elements of the sentence
to preserve isomorphism between the gist of the eventual sentence and whatever
state-of-affairs in the world it denoted. The gist is in a neural modality, unconscious
but intuitable (with some difficulty). A possible form of the gist is an amalgam of key
semantic factors of the two nouns and the verb with the argument structure of the
verb remaining intact (something abstract (but forceful) does something unpleasant
to someone (immature), a rather unusual situation) and with the indefinite modality
(may) and tense embedded in the gist. It thus has a composite gestaltic form that
also incorporates structural features of the sentence such as semantic factors, case,
tense and the grammatical category, plus the argument structure and atelicity of the
verb. There are no words or syntax as such at this level but the gist is structurally
pre-disposed to accommodate both in whatever language the user speaks.

This accommodation occurs at the functional level where the speaker's
language comes into play. Words with their full factoral and gestaltic forms plus their
phonology flow from the mental lexicon bringing with them their kernel attributes that
enable them to combine with the other words. These are their grammatical
categories, their natural combinatory valencies and in the case of the verb, its
transitivity and dual argument structure. At this stage the words all possess their full
phonological, factoral and gestaltic form and their access to encyclopedic meaning.
At the final positional level the language's specific conventions of word order and
agreement provide the final form of the sentence complete with its appropriate
prosodic structure. We have, at last, ‘Sincerity may frighten the boy'. We have traced
its emergence from the ‘deep' neural structure of the gist to the ‘surface' structure of
one spoken language, from neurons to an intelligible sentence. In another language
the structure would have quite different features of morphology, word order and so on.

This is a top-down construction process. There is no need or room for Chomskyan Merge.

This might be regarded as another just-so story but it has, I propose, the benefit of
some vital, new, empirically-based information and it is, I believe, a fuller and more
credible story than it has previously been possible to tell. It is a further illustration of
the power of the previously unknown property of language, the ability of the language
faculty to convert a neural code into a linguistic code in order to represent worldly
states of affairs.

15.5 A new model for the generation of sentences.
The generation of sentences in thought and speech is the primary aim of language.
In order to give more focus to the discussion in the previous section it is useful to
compare the Minimalist process for sentence formation with that inherent in the
approach used in the previous sub-section. There is not scope in this paper to
describe the mechanisms of the Minimalist approach more fully than I already have,
but a schematic description of the factor/gestalt-based approach (FG) can be
compared in more detail by anyone familiar with Minimalism.

The FG process of sentence formation begins with a gist, a neural/mental
formation that motivates the whole process of sentence formation and that contains
the essential material for the construction of the phrase or sentence in a seminal nonlinguistic
form. (I will speak in terms of sentences, but complex, long sentences are
generally contructed in more than one bite.) These materials exist in a
factoral/gestaltic form i.e. in the somatotopic language of the neurons, not in a verbal
form. They comprise the semantic gist of the sentence-to-be-formed, including
possibly tense, case, modality, active/passive form and the like, but no phonological
features. The gist is a template for the formation of the sentence but it lacks words,
word order and linearization. It can be likened to a folded enzyme. It is the necessary
semantic aspect of the sentence existing in a form that is prior to any language, a
universal abstract form. In effect it is a macro-gestalt with a composite factoral DNA,
but it has the capacity to operate as a template for the construction of the final
semantic form with the words and word order of the speaker's language. So what
determines the final form? Where do the necessary elements of syntax come from?

The process that arises from the FG model is entirely different from the
Chomskyan bottom-up process using Minimilism's Merge iteratively, operating on the
compatability of ‘features', mainly of a syntactic type, of neighboring words, as the
driving syntactic element. In the FG model the key to the ‘transformation' from a
neural gist to a sentence in the speaker'' language is the availability of words from
the speaker's mental lexicon. Each word is a package with its phonological and
semantic form, inherent grammatical category, syntactic features, factoral structure,
gestalt (its idiosyncratic DNA) and encyclopedic meaning. These have a powerful
influence on the combination of words. The other key influence on sentence
formation is the availability in speakers' linguistic memories of a multiplicity of learned
‘construction templates'. These are known conventional patterns of word order and
grammatical construction that language users assimilate by exposure to them in the
‘envelope of language' within which they live from birth until their use of language
fully matures.

The ‘transformation' of the gist into a sentence thus has two aspects,
lexicalization and the imposition of an intelligible and acceptable ‘shape' on the
sentence according to the conventions of the speaker's language. In the Vigliocco et
al paper these were viewed as separate stages. Under the FG model they appear to
operate as a continuous process. The first part is the crystallization of words with
their full phonological, grammatical and semantic forms under the influence of the
‘unfolding' of their composite factoral DNA in the gist that effects the matching with
appropriate words from the mental lexicon. The second part is the influence of
learned conventional constructions that provide templates for the final form of the
sentence. This model presupposes the existence in speakers' memories of the many
construction templates that facilitate the construction of well-formed sentences in any
language.

There are two powerful empirical aspects that provide the process I have described
with a degree of credibility: our intuitive access to gists and to the factoral structure of
word meaning. With a little introspective effort and practice we can view gists on the
fly. We can do this better in the process of composing written material than in
speaking. Even then, intuiting them before they take on verbal form is not easy. In
practice many gists are stillborn. They are recognized as non-viable and rejected and
possibly replaced by a viable form of gist. This may happen more than once in the
composition of sentences. Gists also exist in another form, the residual memory of
sentences. We can rarely recall sentences we hear or read fully beyond a few
seconds after experiencing them, but we know their gists, more or less.

As I have noted, we also have intuitive access to the semantic factors. This
access is much easier and more reliable now that the factors have seen the light of
day. There are many examples in the literature of linguists groping for appropriate
semantic features that are now readily available in the form of the semantic factors. I
propose that knowledge of the factors and of gists combined with the role of gestalts,
encyclopedic meaning and construction templates enables us to identify a much
simpler process for the construction of sentences than those devised by the
generative grammarians over the years, including the so-called minimalized version.

Chomsky and his colleagues claim to have identified in Universal Grammar the
lineaments of the innate ‘language organ' that is responsible for the generation of
sentences. This is the warrant for their claim that they are studying language as a
‘biolinguistic' science. The Universal Semantics of the semantic factors with its
numerous ramifications calls this claim into question. The empirical credentials of the
model of sentence construction that I have outlined raises serious doubts about the
Minimilist enterprise in its current form. A major merit of the FG model, apart from its
sound empirical basis is its simplicity, efficiency and elegance, qualities that are
frequently proclaimed for the minimalist approach. If there is such a thing as a
‘language organ', the way the factors form a bridge between the neural instantiation
of word meaning and the semantic content of words as we use them, is surely the
better candidate.

15.6 The primality of words in language.
Generative Grammar has come to give more importance to the lexicon than it did in
its early phase, but its syntactocentricity still runs very deep. Most Chomskyans still
subscribe to the autonomy of syntax hypothesis. On the other hand, I have
emphasized the role of words and that their meanings carry the necessary
requirements to effect much of the combinatorial operation that is central to the
formation of sentences. As we have seen this position is deeply at odds with the
Minimilist Program in which Merge and recursion are held to be key elements of UG
that perform this function. In this respect and in other aspects of my criticism of the
Program I would seem to be in an isolated and vulnerable position. I was not aware
of any other linguist who attached such a central function to words until I recently
became aware of a well-established European scholar with some similar ideas.

Jan Koster is a prominent Dutch linguist with a distinguished career in
linguistics of the Chomskyan variety. He has spent most of his life teaching and
writing about various aspects of Generative Grammar as it has evolved over the last
40 years but he has also become one of its most ardent critics. He is intimate with
the extensive literature of generative grammar but as a European he has a somewhat
different academic genealogy fom his American counterparts. This has influenced his
ideas to some degree. He has long adopted a critical position on some aspects of
generative grammar, and the Minimalist Program in particular, while still being
supportive of many of its technical achievements. In recent years he has produced a
series of papers that potentially have profound implications for Chomskyan grammar
as it is usually presented. In particular he denies the hegemony of syntax and
elevates the word to a central position in the combinatory mechanisms of language.
He also rejects both the generativists' concept of biolinguistics, the assumption of
generativists that UG and the Merge function have a genetic basis, and the
‘internalist' approach to language that is integral to Generative Grammar. These
contentions attack fiercely argued foundational aspects of Chomskyan thought.

Koster is a lexicalist, that is he assumes that words have a central position in
the operation of language. The meanings of words contain ‘templates' that are key
influences on the combinations of words in sentences in his view. ‘Syntactic
structures are properties of words', and from the same paper (Koster 2007), ‘the
most important characteristic of words of real human languages is that their potential
syntactic environments are among their properties in abstraction from any rules or
other word-independent computational devices'. My earlier assertions that word
meanings are endowed with ‘valencies' that are partly derived from the semantic
factors is highly consistent with Koster's position. He speaks of valencies but it is
evident that he is referring mainly to the syntactic valencies of words, argument
structure and grammatical category. He is not aware of the semantic factors except
perhaps in vague intuitive terms. In fact, like Chomsky, he seems to assume that
what he calls ‘the Platonic residue' of words, their universal idiosyncratic semantic
features as types that distinguish them from the tokens of direct experience, is
‘completely beyond the scope of naturalistic inquiry as the physical world is populated
by particulars (tokens) not by universals (types)' (ibid.). From my perspective the
semantic factor configurations associated with words plus their gestalts appear to be
precisely what constitutes this ‘Platonic residue', and they are eminently within the
scope of science in my scenario. This recognition strengthens Koster's argument that
the valencies of words constitute a ‘template structure', for the combination of words
in sentences thus making Chomskyan Merge redundant.

Koster bolsters this position with a number of other arguments for example: (i)
that the meanings of words are human artifacts that have ‘agentive functionality' in
that they are referential and communicational tools for use by language users, a
public/private function, rather than purely internal adjuncts to syntax as is assumed
by generative grammarians; (ii) that ‘words must be associated with coded
information in the brain, not with meaning, because all known physical structures are
without inherent meaning'. Meaning can only be created by someone processing and
interpreting data in the neural informational code.

These and other arguments of Kosters, if they are sustained, have devastating
implications for two pillars of Chomskyan thought, ‘internalism' and the ‘biolinguistics
hypothesis'. Koster does not deny the existence of innate aspects of language or that
language is firmly rooted in biology. I have discussed internalism above.
Biolinguistics has two main aims in the hands of generative grammarians: to discover
if fundamental principles of language can be correlated with natural laws that may
also apply to other fields of biology, and to establish whether, as they hypothesize,
language is grounded on late (very late in evolutionary time) genetically embedded
neurological changes that underpin key elements of Universal Grammar. I do not
have the scope to discuss this controversial matter here beyond noting that Koster
holds that, although there are important necessary connections between biology and
language, these do not imply that a specific genetic change of some kind was
necessary to support syntax and the Merge function that, as we have seen he
attributes, as I do, to words and not to syntax.

The semantic factors provide strong support for Koster's lexicalism because
their structural function is fundamental to the syntactic ‘template' that is inherent in all
words. It underpins words' grammatical categoriality and the argument structure of
verbs. This has potentially grave consequences for the Chomskyan position because
it invalidates Merge, the current mainstay of UG. Lexicalism postulates that the
combinatorial power of language lies in words and is only syntactic in a derived
sense. Syntax is thus shorn of its prime supposed function and is left with a range of
lesser ones. Word meaning, not syntax is the autonomous, foundational, generative
component of language, probably with the support of an intrinsically conventional
tendency of language to form constructions that may account for some of the other
territory claimed by UG. With the factors and gestalts at its centre the word has
arguably been the driving force in language from its emergent beginnings, and is now
in the ontogenetic process of childrens' language learning and in the formation of
every sentence we think, speak or write.

Generative Grammar is severely deflated in this scenario, it seems. The
importance of avoiding the threat that has always been posed by lexicalism may
account for some of the excessive claims for the role of syntax and its consistent
subsumption of semantics. The generativists have been extraordinarily inventive and
productive in building a defensive wall around their territory. The outcome to this
debate cannot be presupposed. Lexicalism needs much more resources than it
currently has. Hopefully these will be recruited over time by its intrinsic appeal as an
alternative to the Chomskyan approach to language.

15.7 Summing up.
There seems to be a very large gulf between Generative Grammar that invests so
much in the functions of syntax and an innate UG, and the approach developed in
this paper that views the semantics of words with their dimensional structure as the
central driving force in language. There is a curious aporia in the Chomskyan position
in relation to semantic features. Generativists view the syntactic features of words as
a key part of the language mechanism. Chomsky often refers to semantic features
more generally but rarely in very specific terms. He often admits to puzzlement as to
the content of word and concept meaning and on several occasions avers that it is
beyond understanding (especially in Chomsky/McGilvray 2012). Words are atomic
and their meanings are innate and they are the source of information that is used
both in the mental computation of sentences and in the interpretation that occurs at
the end of the process. There is much that is obscure and inconclusive in this
thinking which he recognizes in the term 'the Minimalist Program'. The semantic
factors and recognition of gestalts seems to have a great deal to contribute by way of
a new understanding of the content of word meaning, but this will be at great cost to
generative grammar if a lexicalist approach is valid.

The key flaws in Chomskyan Generative Grammar in the light of this paper's
approach are:

........1. Its hypothesis of an innate Universal Grammar and the associated Poverty of
............Stimulus hypothesis both of which lack empirical foundations that cannot be
............strongly contested by alternative hypotheses.

........2. The dependence of the Chomskyan hypotheses on dubious and unsupported
............evolutionary speculations.

........3. Generative Grammar's failure to recognize the arguably empirically wellfounded
............constitution of word meaning that is evident in the 3-part model and
............especially in the semantic factors

........4. Its failure to recognize the natural ‘bio-logic' that is inherent in the semantic
............factors or the influence of commonsense logic in determining the Logical Form
............of sentences.

........5. Its lack of recognition of a non-linguistic Language of Thought that operates
............combinatorially in pre-linguistic thought, as well as in the generation of the
............logically coherent gists that precede the construction of sentences.

........6. Its Rationalist ideology that prioritizes the operation of a speculative innate
............Universal Grammar that is manifested in the supposedly autonomous logical
............structure of sentences over the semantic element of language that can now be
............viewed as operating in the form of the Universal Semantics of words with their
............inherent combinatorial power in association with learned conventional
............templates of phrase and sentence construction with their range of
............syntactic functions.

........7. Chomsky's refusal to acknowledge the intrinsic natural referentiality and
............intentionality that is inherent in the computation of word meaning and is
............manifested in the semantic factors operating as the DNA of language.

Given that these are the views of an unknown language researcher without academic
credentials (albeit with some parallelism in the work of Jan Koster) critics may be
inclined to reject them out of hand, but this will require an alternative account of the
semantic factors and their function in language and, indeed, of the 3-part model.
There are undoubtedly weaknesses and flaws in my arguments and hypotheses that
have been developed without the benefit of peer review. Hopefully these are not fatal.
Much needs to be done in seeking to find a place in linguistic thought for the
approach described in this paper. I am hopeful that this website, in conjunction with
the publication of my small new book on the remarkable deep similarity between the
words of English and of Maori, will begin to achieve some recognition among linguists
for the approach I have outlined. In addition I will continue to endeavour to convince
some neurophysiologists that the function of gestalts and configurations of semantic
factors in word meaning may be able to assist them in interpreting the data they
obtain from brain mapping. At this point I can only speculate on how this might play
out. On the basis of my past experience it is still a formidable task.


annotated contents




16. Summary and conclusions.

This paper has travelled a great distance from the discovery of a previously
unrecognized association between the sounds and meanings of words. I have
described a set of lexical semantic primitives, the semantic factors that have three
properties that appear to be centrally important for understanding language. First,
they provide a previously unrecognized bridge between the neurons of the brain and
the language faculty. Secondly, they constitute the necessary structure of the
meanings of words of all languages in a manner that is arguably grounded in a
biological dimensionality that is common to all forms of cognition. Thirdly they have a
powerful influence in the construction of sentences.

These are very large claims. The factors were identified intuitively in the
meanings of the core words of languages in a way that is sufficiently open to scrutiny
to provide a high level of empirical accountability. The paper proposes that the set of
factoral dimensions also applies across the whole domain of biology. They operate
as the dimensions of the space of the interaction of organisms and their
environments. They are thus a foundational element of life in all its forms. This is ah
astonishing conclusion, to say the least. It highlights the factors' foundational function
in language. To someone familiar with the use of the factors in the analysis of word
meaning, their biological function appears to be the natural basis of their function in
forming the necessary structure of meaning in language as a recently-evolved
modality of organismal interaction with the environment. The wider biological
functions of the factoral dimensions may be difficult to prove but thought experiments
to test its validity in relation to any biological species can support its credibility to a
surprising degree.

The ultimate test of the semantic factors operating in word meaning may lie
with neuroscience. Much has been learnt but current technologies and methods have
yet to cast much light on how the content of word meaning is represented neurally at
a fine-grained level. If I am correct in my conclusion that the factoral and gestaltic
content of word meaning is a bridge between the neural computation of language and
language as we use it, the model would seem to provide an unprecedented insight
for neuroscientists into how word meaning may be investigated.

This paper has reached some conclusions that are very surprising and that must
remain open to question because they are derived from a feature of language that
has not previously been identified (extensive sound/meaning iconicity in some
languages) and that itself is at present based on one person's intuitive interpretations
of lexical data. This may appear to have the makings of a house of cards. On all
counts the basic data needs to be empirically sound and the interpretations credible.
This aspect has been problematic in terms of gaining support from other scholars.
There are several possible reasons for this: claims of association of sound and
meaning in words have been suspect since Plato's surmises about such linkages in
his dialogue ‘Cratylus'; the important linguistic principle of the arbitrariness of the sign
seems to preclude anything more than minimal associations; sound symbolism has
not previously been a source of significant insights into language; intuitive data are
regarded as inherently problematic. In such circumstances it is difficult to convince
scholars that a new, central aspect of language can be found in the associations of
sound and meaning in words. In order to further support the case I have made I will
set out a range of interconnected matters that, I believe, provide robust support for
the soundness of the data on which my conclusions rest.

........1. The sound/meaning iconicity that revealed the factors in English is based on
............its very high predominance in the core words in most sections of the
............dictionary.

........2. There is considerable evidence of this form of iconicity in other languages
............such as Maori.

........3. Phonaesthesia is a well-known phenomenon in English that has some marked
............parallels with the semantic factors and provides evidence of their presence.

........4. The interpretations of the iconicity data are open to testing and will, I believe,
............withstand scrutiny.

........5. The set of semantic factors derived from the interpretations has a high degree
............of coherence and systematicity for the function ascribed to them, providing the
............necessary structure of the meanings of words in all languages.

........6. The factors all have the characteristics required by their postulated function in
............that they have a consistent degree of abstraction and appear to possess the
............comprehensiveness needed for their large task.

........7. They are highly effective, in association with the other two elements of word
............meaning in the proposed model, in the analysis of the meanings of words of all
............
types
at the structural level.

........8. Once one is familiar with the semantic factors their identification in the
............meaning of core words is perspicuous.

........9. The factors manifest a property that linguists have frequently assumed to be a
............property of semantic features, polarity.

.......10. An extension of the domain of the factors to human cognition and behavior, as
............their original human domain, is credible and rich in explanation, as is their
............hypothesized ultimate origin in the regulation of organisms' interaction with the
............environment across the whole realm of biology.

.......11. The factors appear to have considerable potential to be useful in clarifying
............such unresolved matters as the origin of language, the early emergence of
............language in children, the nature of thought and the relative roles of word
............meaning and syntax in the structure of language.

......12. The factors provide a powerful basis for a critique of Generative Grammar in
............its various forms.

This is a formidable armoury of evidential support that compensates to a large
degree, I believe, for the lack of more concrete neurophysiological evidence and the
current lack of support from academia. From my perspective it justifies the metaphors
I have used to describe the nature and significance of the factors: alphabet of
meaning, source code from which the computational system of language is built, the
DNA of language and a Rosetta Stone that serendipitously reveals the long-sought
featural structure of word meaning.

It is idle to speculate what the course of investigation of word meaning would have
been like if the semantic factors had been discovered 30 or 40 years ago but it would
seem likely to have been somewhat different because it has revealed the power of
words in the larger constructions of language. As we have seen, the factors are not
difficult to identify. The failure to do so was in part a consequence of five taboos: the
belief that Saussure's arbitrariness principle was virtually absolute; the belief that
sound symbolism is a dead-end in linguistic terms; the supposed inadequacy of the
introspective investigation of meaning; the lingering effect of Behaviorism's antipathy
to mental images that culminated in the ‘imagery debate' of the 1970s and 80s; and
an associated prejudice against similarity and iconicity that stemmed in part from
philosophical arguments (Goodman 1957, Sloman and Rips 1998). Among other
things this paper is an attempt to exorcise these taboos.

The investigations I have undertaken and their results as outlined here are the first
steps in exploring the make-up of word meaning in a new way and evaluating their
implications for the investigation of language. I hope that other researchers will take
up the challenge of building on this beginning.

14.10.13


annotated contents




References.


Barsalou, L. W. 1999. Perceptual Symbol Systems. Behavioural and Brain Sciences
............22, 577–660.

Barsalou, L. W. 2007. Grounded cognition. Annual Review of Psychology, Volume
............59, 2008.

Barton, R. A. 2013. A Skeptics Guide to the Mind: What meuroscience can and
............cannot tell us about ourselves
. St Martins Press.

Bergen, B.K. 2004. The psychological reality of the phonaestheme. Language 80(2)

Bierwisch, Manfred. 1971. On classifying semantic features? In Semantics. Ed, D.D
............Steinberg and L. A. Jacobovits. Cambridge.

Binder, J. R., Desai, R. H , Graves, W. W., Conant, L. L. (2009) Where Is the
............Semantic System? A Critical Review and Meta-Analysis of 120 Functional
............Neuroimaging Studies
. Cerebral Cortex 19/12.

Binder, J. R. McKiernan K. A. Possing, E. T. Medler, D. A. (2005). Distinct brain
............systems for processing concrete and abstract concepts
. Journal of Cognitive
............Neuroscience. 17(6).

Chang, K. K., Mitchell, T., and Just, M. A. 2011. Quantitative modeling of the neural
............representation of objects: How semantic feature norms can account for fMRI
............activation
. Neuroimage 56 716–727.

Chomsky, N. 1965. Aspects of the Theory of Syntax. The M.I.T. Press.

Chomsky, N. 1995. The Minimilist Program. Massachussets Institute of Technology.

Chomsky, N. 2000. New Horizons in the Study of Mind and Language. Cambridge
............University Press.

Chomsky, N. and McGilvray, J. 2012. The Science of Language. Interviews with
............James McGilvray
. Cambridge University Press.

Crain, S. 2012. The Emergence of Meaning. Cambridge University Press.

Fodor, J. 1975. The Language of Thought: A Philosophical Study of Cognitive
............Science
. Hassocks: Harvester Press.

Fodor, J, 2008. LOT2: The Language of Thought Revisited. Oxford University Press.

Gaskell, H. G. (Ed.) 2009. The Oxford Handbook of Psycholinguistics. Oxford
............University Press.

Givon, T. 2009. The Genesis of Semantic Complexity: Diachrony, ontogeny, neurocognition,
............evolution
. John Benjamin Publishing Company.

Goddard, C. 1998. Semantic Analysis: A Practical Introduction. Oxford University Press.

Goddard, C and Wierzbicka, a. (Eds) 1994. Semantic and Lexical Universals. John
............Benjamins.

Goodman, N. 1972. Seven Strictures on similarity. In: Goodman, N. (ed.) Problems
............and projects. Bobbs-Merrill, New York.

Hinton, L., Nichols, J. and Ohala, L. 1994. Sound Symbolism. Cambridge University Press.

Jackendoff, R. 2004. Foundations of Language: Brain, Meaning, Grammar,
............Evolution
. Oxford University Press.

Johnson-Laird, P. N. 1983. Mental Models. Towards a Cognitive Science of
............Language, Inference and Consciousness
. Cambridge University Press.

Johnson, M. 2007. The Meaning of the Body. University of Chicago Press.

Just, M. A., Cherkassky, V.L., Aryal, S., Mitchell, T. M. 2010. A Neurosemantic
............Theory of Concrete Noun Representation Based on Underlying Brain Codes
.
............PLoS ONE Volume 5 Issue 1.

Hinton, L., Nichols, J. and Ohala, L. 1994. Sound Symbolism. Cambridge University Press.

Hinzen, W and Poeppel, D (eds.) 2011. Language and Cognitive Processes.
............Psychology Press, Oxford.

Katz, J and Fodor, J. (1963) The structure of a semantic theory. Language 39 170 – 210.

Kosslyn, S. M. 1994. Image and Brain: The Resolution of the Imagery Debate.
............Cambridge MA: MIT Press.

Kousta, S-T., Vigliocco, G., Vinson, D. P., Andrews, M., Del Campo, E. 2011. The
............representation of abstract words: Why meaning matters
. Journal of
............Experimental Psychology. 02/2011.

Kuchinke, L., Jacobs, A.M,. Grubich, C., Vo, M. L., Conrad, M., Herrman, M. 2005.
............Incidental effects of emotional valence in single word processing: an fMRI
............study. Neuroimage. 28(4):1022-32.

Lloyd, T. 2003. Fossils of Language: Primal Meanings in the Sounds of Words.
............Steele Roberts, Wellington, New Zealand.

Lloyd, T. 2008. The Sound of Meaning. (Unpublished paper).

Lloyd, T. 2009. The Alphabet of Meaning. (Unpublished paper)

Lloyd, T. 2009. A New Biological Grounding for Cognition and Language.
............(Unpublished paper).

Lloyd, T. 2013. How to find a taniwha: a deep connection beween the meanings of
............Maori and English words
. Steele Roberts, Wellington.
............McRae, K., Cree, G. S., Seidenberg...

Seidenberg, M. S., McNorgan, C. 2005. Semantic feature production norms for a
............large set of living and nonliving things
. Behavior Research Methods,
............Instruments and Computers. 37(4) 547–559.

Mitchell, T. M., Shinkareva, S. V., Carlson, A., Chiang, K.-M., Malave, V. L., Mason,
............R. A., Just, M. A. 2003. Predicting human brain activity associated with the
............meanings of nouns
. Science Vol. 320.

Moss, H. E., Tyler, L. K. and Taylor, K. I. 2007. Conceptual structure. In Gaskell
............(Ed.), 2009.

Nanny, M and Fischer, O. (Eds.) 1999. Form Miming Meaning. John Benjamins.

Ohala, J. 1994. The biological basis of sound symbolism. In Hinton, Nichols and Ohala.

Poeppel, David and Embick, David. 2005. "Defining the Relation Between Linguistics
............and Neuroscience
". In Anne Cutler. Twenty-First Century Psycholinguistics:
............Four Cornerstones. Lawrence Erlbaum.

Pexman, P. M., Hargreaves, I. S., Edwards, J. D., Henry, L. C., Goodyear, B. G
............2007. Neural correlates of concreteness in semantic categorization. Journal of
............Cognitive Neuroscience 19(8):1407-19.

Pustejovsky, J. 1995. The Generative Lexicon. Cambridge, Mass.: MIT Press.

Pylyshyn, Z. 1984. Computation and Cognition: Toward a Foundation for Cognitive
............Science
. MIT Press.

Rosch, E. 1978. Principles of categorization. Pp. 27–48 in Rosch, E. & Lloyd, B.B.
............(eds), Cognition and Categorization, Lawrence Erlbaum Associates.

Rosch, E and Mervis, C. B. 1975. Family resemblances: Studies in the internal
............structures of categories
. Cognitive Psychology: 573–605.

Sabevitz, D. S., Madler, D. A., Seidenberg, M., Binder, J. R. 2005. Modulation of the
............semantic system by word imageability
. Neuroimage. Aug 1;27(1):188-200.

Satel, S., and Littlefield, S. O. 2013. Brainwashed: The Seductive Appeal of Mindless
............Neuroscience
. Basic Books.

Saussure, F. de 1995. Course in General Linguistics. Bally, C and Sechahaye, A,
............(Eds.) New York Philosophical Library.

Simmons, P. and Young, D. 2010. Nerve Cells and Animal Behaviour. Cambridge
............University Press.

Sloman, S. A and Rips, L. J. 1998. Similarity and Symbols in Human Thinking.
............Elsevier Science Publications, B. V., Amsterdam.

Steinberg, D. D. and Jakobovits, L. A, 1971. Semantics. An Interdisciplinary Reader
............in Philosophy, Linguistics and Psychology
. CambridgeUniversity Press.

Taylor, K.I, Devereux, B. J., Tyler, L. K. 2011. Conceptual structure: Towards an
............integrated neurocognitive account
. In Hinzen and Poeppel.

Vigliocco, G. 2007. Semantic Representation. In The Oxford Handbook of
............Psycholinguistics
. Ed M. J Gaskell.

Vigliocco, G., Vinson, D.P, Lewis, W. & Garrett, M.F. (2004). Representing the
............meanings of object and action words: The featural and unitary semantic space
............hypothesis
. Cognitive Psychology, 48, 422-488

Weinrich, U. 1963. On the semantic structure of language. In Universals of Language
............ed. Joseph H. Greenberg. 1963

Weinrich, U. 1966. Explorations in semantic theory In Semantics: An Interdisciplinary
............Reader
. Cambridge.

Wilson-Mendenhall, C.D., Barrett, L.F., & Barsalou, L.W. 2013. Neural evidence that
............human emotions share core affective properties
. Psychological Science, 24,
............947-956.


annotated contents




Annex 1: A full description of the semantic factors.


The descriptive factors.

(Factor poles are in brackets).

m Materiality • 1. Materiality, substance; 2. association with materiality; 3. extendedness,
horizontality, laterality. 4. magnitude, multiplicity, plurality; 5. measurement; 6. manipulation of
materials, artifactuality; (7. immateriality, the abstract, the mental).

p Particularity • 1. Particularity, discreteness, specificity, singularity; 2. exclusivity; and the following
character of things: 3. seriality, severality, repetition; verticality; long and thin; flat and bounded;
small and compact; (4. generality, multiplicity, durativity).

f Surface • 1. Surface; 2. flatness; 3. the manipulation of surfaces; 4. physical, social, familial and
personal closeness or intimacy; (5. excessively detailed; 6. insubstantiality, unreliability, falseness).

b Bodily roundedness, convexity, animacy • 1.Bodily roundedness, bodiliness, animality; 2.
convexity; 3. animacy, animality, natural kind; (4. violence, destruction; 5. inanimacy, artifactuality).

n Contraction • 1. Contraction, compression, reduction, focus; 2. attraction, self-directed action,
intransitivity; 3. negation. (4. dispersion, emission, separation, release, expression; 5. externallydirected
action, transitivity).

y Extension • 1. Extension, stretching, linearity; 2. continuous, ongoing, duration, temporality; (3.
immediateness, non-duration).

l Display • 1. Display, manifestness, conspicuous visibility, salience, expression; 2. illumination; (3.
concealment, obscurity, latency).

o Openness • 1. Spatial openness, spatiality 2. personal openness, mental space; 3. possibility,
opportunity; (4. concealment, confinement, absence).

a Action, largeness • 1. Action, activity, process; 2. largeness, enlargement;

i Smallness, interiority, identity • 1. Smallness; 2. interiority, enclosure; 3. Identity, essentiality; 4.
having an abstract or mental quality.

r Intensity of effort • 1. Effort or energy of various intensities; 2. activity, movement, event, process,
state; (3. effortlessness, inactivity, rest, spontaneity).

k Physical intensity • 1. Intensity of quality, distinctiveness, typicality; 2. the prominence of parts of
things, angularity; 3. intensity, unusual form, complexity, refinement and intricacy of things and
actions; (4. featurelessness, blandness).

2. The affective factors.

d Positiveness / negativeness • 1. Bright, happy, good, pleasant, favourable aspects of the world;
attraction; 2. dark, sad, bad, irritating, unfavourable aspects; aversion.

v, z Fullness / emptiness • 1.Excellence, salience; 2. fullness of intrinsic values such as
positiveness, forcefulness, presence, mindfulness; 3. physical or figurative fullness; (4. emptiness,
insubstantiality, lack, absence, negativeness; 5. physical or figurative emptiness that is also a capacity
to contain; containment).

s Body and affects • 1. Bodily-felt affect; 2. bodily and psychological actions and states, personal
states; 3. sensations; 4. associated with the body. (5. affective neutrality).

t Tactility • 1. Touch, tactility, contact; 2. sensation; 3. pointing, deixis; 4. uprightness.

g Abundance, generosity • 1. Abundance, availability; 2. generosity, bountifulness, goodwill, care; 3.
productivity, effectiveness, affordance; (4. paucity, deprivation; 5. seeking, grasping).

h Possession, self-interest • 1. Possession, acquisition, territoriality; 2. want, desire, appetite; 3.
self-interest; 4. personhood, subjectivity; (5. loss, detriment, hostility, threat,).

j Personal energy, humour • . Personal energy, motivation; 2. humour, happiness;

q Uncertainty • 1. Uncertainty, indeterminacy; (2. definitiveness, authority).

u Displacement • 1. Displacement, dislocation, disorder, being out of kilter, negation; 2. wrongness,
uselessness, discomfort; 3. disparagement, (4. correctness, satisfaction, comfort, normality, order and
utility).

w Existential value • 1. The existential significance and value of the everyday.

3. The taxonomic factors.

c, e The lifeworlds • 1. The human, social and family lifeworld and 2. the purely physical lifeworld.

.


annotated contents




Annex 2: The full set of initial phonaesthemes.

This annex contains all the word-initial phonaesthemes in English. I will paraphrase
their often somewhat obscure, sometimes inconsistent, always distinctive but difficult
to verbalise, meanings; list some of the main instances of each phonaestheme; and I
will indicate the semantic factors that appear to be associated with the individual
phonemes of the clusters as discussed in Section 4. The word listings are intended
only to illustrate the semantic range of phonaesthemes. A notable feature is that, in
most phonaesthemes some words are not a neat fit with the core theme due to the
way phonaesthemes develop. It is as though phonaesthemic meaning is constantly
outgrowing its original narrow confinement in a specific concept. The purpose of this
data is to provide a basis for some discussion in the paper on the nature of
phonaesthesia, what motivates its development, and the role of the phoneme sound
symbolism in this phenomenon.

bl-: defective appearance: black, blaze, blemish, blind, blink, blot, blue, blur, blush:
......surface; display, manifestness.

br-: violence, intensity: brace (v.), brash, brave, brawl, break, bruise, brunt, brute:
......violence; intensity of energy.

cl- : in a tight grip: clam, clamp, clasp, claw, clench, climb, cling, close (v. and adj.),
......cluster, clutch: physical intensity; display (polar).

cr- : having a bent appearance: cramp, crawl, creep, cringe, cripple, crouch, crush:
physical intensity ; intensifier.

dr-: protracted negative action: drag, draw, drawl, dribble, drift, drip, droop, drown:
......negativity; intensifier.

dw-: diminishment: dwarf, dwell, dwindle: positive/negative?; ?

fl-: prominent surface: flag (n.), flame, flash, flat, flesh, fling, flip, float, flood, floor,
......flower, fly: surface; display.

fr-: distinctive appearance: frantic, freckle, freeze, fresh, friend, fright, front, froth,
......frown, fry (v.): surface; intensifier.

gl-: emission of light: glare, gleam, glint, glimmer, glisten, glitter, gloss, glory, glow:
......abundance; illumination.

gn-: chewed: gnarled, gnash, gnaw: physical intensity (k); contraction.

gr-: deprivation, seeking: grab, grasp, greed, grief, grip, grope, grovel, growl, grudge,
.......grumble: negative pole of abundance; intensifier.

kn-: distinctive physicality: knee, knit, knock, knot, know, knuckle: physical intensity;
.......compression, contraction
.

pl-: bounded displayed surface: place, plan, plane, plank, plaster (v.), plate, plume,
......plump: particularity; display.

pr-: focussed effort: practice, prance, prank, pray, preen, press, prime, print (v.),
.......probe, prove, prowl, pry: particularity; intensifier.

sc-: harsh action: scald, scale (v.), scamper, scarce, scare, scathe, scatter, scoff,
.......scold, scorch, scorn, scour, scuff: bodily action; physical intensity.

scr-: harsh, extended bodily action: scramble, scrape, scratch, scream, screw, scrub:
........bodily action; physical intensity; intensifier;

shr-: reduction from normal state: shred, shrewd, shriek, shrill, shrink, shrivel, shroud,
........shrub, shrug: bodily action; intensifier.

sk-: effortless movement: skate, skew, skid, skill, skim, skimp, skip, skirmish:, bodily
.......action;
physical intensity (polar sense).

sl-: involuntary action (wetness, slipperiness): slack, slam, slash, slay, sleek, sleep,
......slide, slime, slip, slither, slobber, slosh, slouch, slow, slump: bodily action; display.

sm-: dispersed effect on a surface: smack, smash, smell, smile, smite, smoulder,
........smoothe, smudge: bodily action; material (surface).

sn-: associated with the nose: sneer, sneeze, sniff, snore, snort, snout, snuffle: body,
.......bodily action; contraction, compression
.

sn-: bodily or physical contraction: snack, snare, snatch, snap, sneak, snip, snoop,
.......snow, snuggle: bodily action; contraction, compression.

sp-: visual, particulate: spare, spark, sparse, spasm, speak, speckle, spell, spew,
.......spike, spill, spine, spit, spout, spurt: sensation; particularity.

spl-: unsightly, particulate display across a surface: splash, splatter, splay, split,
........splutter: bodily action; particularity; display.

spr-: dispersive action: sprawl, spray, spread, spring, sprinkle, sprout: bodily action;
........particularity; effort
.

squ-: unconstrained dispersal: squabble, squander, squash, squeal, squeamish,
.........squelch, squirm, squirt: bodily action; uncertainty.

st-: upright and motionless: stab, stack, stake, stalk, stamp, stand, stare, stay, steep,
......stick, stiff, still, stop, stun: bodily action; uprightness.

str-: horizontal extension: straggle, straight, strain, streak, stream, stretch, stride,
........string, strip, stripe, stroke: body, bodily action; tactility; intensity of energy.

sw-: broad mainly lateral movements: swab, swagger, swap, swarm, swat, sway,
........sweep, swerve, swim, swing, swipe, swoop: body, bodily action; ?

thr-: pronounced tactility: thrash, thread, threat, thresh, thrift, thrill, thrive, throb,
........through, throw, thrust: tactility; intensifier.

tr-: extended, laborious action: track(v.), trade, trail, train (v.), tramp, tread, trip (n.),
......trudge, try: tactility, intensity of action.

tw-: tactility, diminution (two-ness): twiddle, twig, twine, twinge, twinkle, twist, twitch,
.......twitter, two: tactility; ?

wr-: laborious bodily action: wrangle, wrap, wreck, wrench, wrestle, wriggle, wring,
........writhe, wrought: ?; intensity of action.

Two general points warrant mention. First, it is remarkable that all the initial
consonant clusters in English arguably have associated phonaesthemes. This attests
to the power and significance of this mechanism in word formation. Second, almost
half of the phonaesthemes begin with /s/. This is a consequence of the direct
association of this phoneme with the body (see Section 7) and the strong connection
all the phonaesthemes have with the body and its senses.




.




annotated contents