A critical assessment of the following dissertation
•
•
Target length: 800 words
Assesses learning outcomes 1-5
In general the following aspects of the dissertation should be addressed in your critique
although these will vary depending upon the dissertation. Feel free to address other points as
you see fit. Note that this assignment is not just a list of yes/no questions but these types of
question should be addressed in your report. The report should also enable the reader to learn
what the dissertation was about and what the main conclusions were.
The Dissertation
•
•
•
•
•
•
•
•
•
•
•
What topic or area of linguistics is being addressed?
Do you think that sufficient background information or context is provided in the
Literature Review/State of the Art?
Is there irrelevant material in the literature review? Does it appear that the author is
including everything he/she has read?
Do the subsequent chapters follow in a natural manner from the literature review? Is
there a research question and what is it?
What research method(s) is used, e.g., quantitative, qualitative, mixed?
If a quantitative method is used, are any statistical methods discussed?
If a qualitative method is used is there an analytical or theoretical framework used for
the analysis of the data?
What results are found?
How do the results relate to the background?
Do the results support the theory?
Approximately what percentage of the dissertation is taken up with: (i) the
Introduction; (ii) the State of the Art; (iii) Results; (iv) Discussion; (iv) Conclusion.
Note that if your dissertation is judged to be one of the best ones, we may well put it on
blackboard (anonymised) for next year’s students to critique in terms of the above
questions.
1
From Thought to Language: Exploring a
Cognitive Approach to Language Production
Abstract
The Cognitive Linguistics approach has rarely contemplated certain aspects of Language
Production. Since the scope of a study involving every facet of this area is rather vast and
diverse, it is best to focus on a single aspect at a time. Hence, this paper deliberates only on the
notion of Lexical Access from the Production point of view. Apart from glancing at the current
theories of Lexical Access, the paper aims to explore how a model based on the tenets laid out
by Cognitive Linguistics, namely the Theory of Access Semantics (also known as LCCM
Theory) as proposed by Evans (2009), might be used to accommodate this phenomenon. The
study thus examines if this system could work as a functional model for Lexical Access, while
suggesting certain modifications for it to fit the required criteria. The main area of interest is
enumerating the factors affecting this process – namely Semantic Specificity, Context, Degree
of Semantic Emphasis, Degree of Articulatory Emphasis and Semantic/Articulatory
Generalisation. Consequently, the paper attempts to offer a theoretical account for a functional
model (described as the CMLC Model) that explains the way in which words are selected for a
specific context – on a monolingual and multilingual/cross-linguistic level. It must be noted that
this is by no means a fully-developed theory, and requires a great deal of further research for the
ideas stated herein to reach fruition at that level.
2
1 CONTENTS
2
3
4
Introduction .......................................................................................................................... 4
2.1
Aim and Objectives ........................................................................................................ 5
2.2
Research question and hypothesis .............................................................................. 10
Theoretical Background ...................................................................................................... 11
3.1
Commitments to and of the Approach ........................................................................ 11
3.2
Subsequence and Cascade/Connectionist Models of Lexical Access ........................... 13
3.2.1
Subsequence Model: Levelt and Schriefers (1987) ............................................. 13
3.2.2
Cascade Model: Dell, et al. (1999)....................................................................... 15
3.3
The LCCM Theory ......................................................................................................... 17
3.4
Synonymy .................................................................................................................... 33
3.5
Equivalence.................................................................................................................. 36
3.6
Equivalence from a Cognitive Linguistics Perspective ................................................. 37
3.7
Entrenchment, Collocations and Salience ................................................................... 39
3.8
Entrenchment and Equivalence – A Debatable Dichotomy ......................................... 42
CMLC – A Functional Lexical Access Model? ....................................................................... 44
4.1
LCCM Upside-Down: CMLC? ........................................................................................ 44
4.2
Factors Affecting Lexical Access in Language Production ........................................... 50
4.3
The CMLC Model: An Overview ................................................................................... 60
5
Conclusion ........................................................................................................................... 64
6
References ........................................................................................................................... 67
7
Appendix A – List of Figures ................................................................................................ 70
3
Acknowledgements
I take this opportunity to thank my family and friends for their constant support
throughout this endeavour. Special thanks to my dissertation supervisor and the
personal tutor for being so patient with me and helping channelise my wayward
thoughts into something productive. I would also like to express my gratitude to my
fellow linguists – Jenny, Sahar, Jordan, Leo and Albert, as well as all the lecturers for
their unending support and constructive input. Most of all, I thank the university for
giving me the chance to learn with and from such bright stars from the world of
Linguistics.
4
2 INTRODUCTION
Language Production is as human a tendency as is visual perception, bipedalism,
culture, emotions or any process involving higher-level cognitive functioning. In recent
views on Language Production and the way in which it occurs, a set of “informationally
encapsulated” units (Fodor, 2008) are described – wherein distinct phonological,
syntactic and semantic components are believed to co-exist and interact.
For reasons explained below, in this study, we choose to follow the tenets laid out by
the relatively new (Evans & Green, 2014) discipline of Cognitive Linguistics – a school
of thought that involves the understanding that language is incorporated into the
cognitive system along with other cognitive mechanisms such as conceptualisation,
recall and recollection, and even perceptual processing. It is a matter of consensus
among Cognitive Linguists that language seems to be inextricably connected to each of
these systems, among others.
Despite the fact that Cognitive Linguistics is now a thriving approach, a preliminary
glance at various books, journals etc. has revealed very little work on the subject from
this perspective except for the preliminary foundation laid by Levelt and Schriefers
(1987), Levelt (1992) and Dell & O’Seaghdha (1992) and those building on it. Apart
from that, the indicated theories seem to overlook the factors affecting the process of
Lexical Access, showing greater focus on defining the process itself. The notion of
“entrenchment” (Schmid, 2007), which is an inherent part of Cognitive Linguistics,
seems to play an important role in this process, but requires a more detailed
investigation of how it works with regards to Lexical Access. As a result of this need to
study factors such as entrenchment, which have their theoretical bases in Cognitive
Linguistics, the paper employs the tenets laid out by this field and uses the point of view
presented by it. In doing so, it also aims at maintaining a streamlined focus in terms of
scope, rather than attempting to explain every single facet of Language Production,
Cognitive Linguistics or even Lexical Access. This dissertation thus seeks to remedy the
issues posed, by exploring simply whether a well-articulated Cognitive Linguistics
approach such as Evans’ Theory of Access Semantics (2009), might be used to explain
the point at hand, while other relevant factors included in the field follow suit.
5
2.1 AIM AND OBJECTIVES
The overall aim of this dissertation is to propose a detailed model of the factors
affecting Lexical Access in Language Production, involving examples of adjectives
and other parts of speech across languages in order to bolster the claims made. The
objectives of the study are as follows:
i) To explore the existing theories of Lexical Access from the Cognitive
Linguistics perspective, and suggest a more comprehensive model for the
same. In particular, to explore Evans’ theory of Access Semantics (Evans,
2009), which proposes the pairing of conceptual profiles with specific lexical
items, from a Language Production point of view.
ii) Subsequently, to study “equivalence” substitution: how the absence of a
particular word to signify a concept can enable the selection of a different
word or phrase that might seem equivalent in some way.
iii) More specifically, to study the phenomenon of entrenchment through the
means of lexical associations, and how they manifest in discourse – whether
spoken or written.
iv) Finally, to describe the semantic and articulatory aspects involved in the
process of Lexical Access, and study how they interact with each other.
We now discuss each of these points individually.
i)
Current Theories of Lexical Access: Levelt and Schriefers (1987) describe a
model which involves the “modular” approach; suggesting that the process
of lexical access starts with the “Conceptualiser,” which “maps a
communicative intention onto a preverbal message” (Levelt & Schriefers,
1987, pg. 397) followed by the “Grammatical Encoder,” which uses this
preverbal message in order to produce the “surface structure.” This is then
carried through to the “Sound Form Encoder,” which “applies the surface
structure to a phonetic and articulatory plane” (Levelt & Schriefers, 1987,
pg. 397) which in turn is produced by the “Articulator.” In a later paper by
Levelt (1993), levels 2 and 3 were combined to form the “Formulator.”
While this forms a basic working premise within which lexical access could
6
occur, we take things one step further in this paper by suggesting the
presence and interaction of a number of dimensions or factors that affect the
process.
On the other hand, there exists a separate theoretical approach to the
process of Lexical Access. Proposed by Dell, et al. (1999), this “cascade”
model suggests that the Grammatical Encoder and Sound Form Encoder
work in a parallel manner, rather than sequentially. This could certainly be
the case, as will be discussed in chapters 3 and 4, and is relevant as one of
the cornerstones for the Factors Affecting Lexical Access to be built around.
ii) Equivalence Substitution: The concept of “equivalence,” as proposed by
Jakobson (1959), shall also be taken into consideration herein, with regards
especially to the aspects of “Context,” “Semantic Specificity” and
“Semantic/Articulatory Generalisation.” This paper shall study the existence
of intra-lingual equivalence (Jakobson, 1959), and the ability of individuals
to substitute one word or phrase for another when required.
iii) The role of Entrenchment in Discourse: A vital aspect when it comes to
lexical access, entrenchment is also a very intrinsic part of the Cognitive
Linguistics school of thought. The extent to which the usage of a particular
word or phrase is entrenched is vital in this scenario, especially when it
comes to “Context” as a factor of lexical access. This shall also be examined
further in the Theoretical Background.
iv) Factors Affecting Lexical Access: As mentioned above, the main aim of this
paper is to suggest a set of factors affecting the process of lexical access,
while also taking into consideration the way in which they interact, and how
they may be related to other cognitive and perceptual processes; the
relationship which forms the basis of Cognitive Linguistics. We suggest the
existence of 5 factors – Semantic Specificity, Degree of Semantic Emphasis,
Degree of Articulatory Emphasis, Context and Semantic/Articulatory
7
Generalisation. These shall be reviewed in detail in chapter 4. However, the
obvious question remains to be answered – why these factors?
The elements mentioned above are broadly divided into Semantic,
Articulatory and Contextual operations. It can be assumed that language is
set within these categories; Lexical Access in spoken language requires
conceptual representation (meaning), articulatory representation (speech)
and contextual awareness (linguistic and extralinguistic framework) in order
to function in an efficient and effective manner.
Given this idea, we propose that the factors involved in conceptual
representation would include the following:
-
Semantic Specificity (narrow conceptual definition) – Not to be
confused with the “semantic specificity hypothesis” suggested by Pine,
et al.. (2010), this relates to the activation of specific parts of the
conceptual system, or as explained in chapters 3 and 4 on the basis of
Evans’ (2009) LCCM Theory, those of the Cognitive Model Profile.
-
Degree of Semantic Emphasis (convention-driven emphasis attached to
a specific word or phrase) – This is based on the notion of “extreme and
scalar adjectives” as proposed by Paradis (2001), wherein Bolinger’s
(1967, pg. 4) statement is cited – “…comparability is a semantic feature
coextensive with ‘having different degrees’ or associated to items which
are ‘susceptible to be laid out on a scale’”
-
Semantic Generalisation (broad conceptual description) – This factor is
related to the notion of the Cognitive Model Profile as per Evans’ (2009)
claim.
The Articulatory representation would involve:
-
Degree of Articulatory Emphasis (emphasis created by prosodic factors)
– This is based on the suggestion that “lexical stress,” “lexical tone,”
and other syllabic and segmental information as “independently
represented dimensions,” Goldrick (2014, pg. 236).
-
Linguistic Context - This point finds basis in Evans’ (2009) LCCM
Theory.
8
-
Articulatory Generalisation (broad phonological similarities) – This is
based on a notion proposed by Dell, et al.. (1999) in their “cascade”
model.
Contextual awareness includes various aspects of the Linguistic and
Extralinguistic Context, as described by Evans in his account of the LCCM
Theory (2009).
These factors have been studied under various terminologies before; in this
paper we attempt to consolidate them all into a single working model to see
how they interact. While we do not suggest that they are the only factors
involved, it must be noted that they are certainly important when
considering Lexical Access, as is substantiated by their respective scholarly
works. Given that most of them fall within the Cognitive Linguistics
approach, and many are based on Evans’ (2009) LCCM Theory, these two
aspects shall become the primary fulcra of the suggested model.
This paper thus reviews the Theory of Lexical Concepts and Cognitive Models
(LCCM Theory, also known as the Theory of Access Semantics) as suggested by Evans
(Evans, 2009), and attempts to use its structure as a fundamental premise on which to
base the claims made. Evans’ theory of Access Semantics creates a plausible argument
for the way the cognitive system functions with reference to the perception of language,
but does not necessarily mention that it is limited to comprehension. That is, it mainly
explores the way the hearer or reader of language would access the conceptual
information within their cognitive system in order to make sense of the items they are
presented with. These items, according to Talmy (2000) (cited in Evans, 2009 pg. 102),
are divided into “closed class” and “open class” categories. Despite it being a theory of
Access Semantics, wherein conceptual information is accessed through words, the
LCCM Theory could be turned ‘upside down,’ to see if it could be expanded to include
the functioning of language production as well. That is, while the current theory deals
with the Language>Cognition process, we shall attempt to see if it works from the
Cognition>Language perspective.
Given these points, it may be assumed that if the system of Access Semantics works
along these lines, then that of Lexical Access would work in the opposite direction.
9
However, in this study, we suggest that while the basic premise might be true, when it
comes to production, a few more factors need to be taken into consideration as a means
of expanding the given theory.
The methodology involved herein consists mainly of the theoretical approach, while
also using supporting information drawn from corpora to substantiate the claims made.
The use of 4 different languages – namely English, Hindi, Marathi and Gujarati, is twofold in its objectives. The first aim is to explore the variety of occurrences and
variations that occur across languages that vary on the scale of similarity. The second is
aimed at bolstering the claims made further, so as to suggest a generic, working theory
on a cross-linguistic level.
It may be noted during the course of the paper that the examples stated involve
mainly adjectives. This is because adjectives seem to provide a wide range of
possibilities when it comes “scalar” and “extreme” (Paradis, 2001) kinds of instances,
so as to create further clarity of functioning in terms of the factors mentioned above.
The theory could, of course, also be applied to other parts of speech, including nouns,
verbs, adverbs, prepositions, etc.
It has been suggested that there are no absolute “synonyms” (Cruse, 2000). With this
point in mind, we suggest the use of the term “equivalents,” as indicated by Jakobson
(1959), in order to discuss the process of substitution and the elements thereof. One can
consider two kinds of “equivalents.” The first category is the “semantic” equivalents,
which would focus mainly on ‘getting the point across through any means necessary,’
without any consideration towards formal or grammatical similarity between the source
and target equivalents. The best example of this would be the use of a phrase in order to
substitute a word, usually seen in spoken discourse. On the other hand, there exist the
“formal” equivalents, which not only include the use of words within the same
grammatical and semantic category, but also ones that may not necessarily carry similar
semantic structure. The best example of this is the use of the word “thing,” or in
informal discourse, even “thingy,” to express the existence of a noun that for some
reason has either slipped the speaker’s mind, or has lesser semantic focus and greater
perceptual focus, as in Hand me the thing. This point shall also be discussed further in
the Theoretical Background.
10
2.2 RESEARCH QUESTION AND HYPOTHESIS
The research questions for this study are as follows:
a) Which aspects of the current theories of Lexical Access in Language
Production can be combined to make a more holistic, cross-linguistic model?
Can the theory of Access Semantics, as proposed by Evans (2009)be used as
a working model of Lexical Access in this scenario?
ii) What is the role of “equivalence” as suggested by Jakobson (1959) in his
theory with regards to the suggested model of Lexical Access?
iii) What role does Entrenchment (Schmid, 2007) play in this process of Lexical
Access?
iv) What are the factors affecting Lexical Access in Language Production?
*It is to be noted that any mention of Lexical Access in this study is made with reference to Language
Production, unless otherwise specified.
The main hypothesis is as follows:
“The factors affecting Lexical Access, which form the crux of the process involved,
are Semantic Specificity, Degree of Semantic Emphasis, Degree of Articulatory
Emphasis, Context and Semantic/Articulatory Generalisation. These factors interact
with aspects such as Entrenchment and Equivalence, and attempt to bring forth a
number of potential lexical units that may be used, which undergo further filtration,
extracting the specific lexical unit that is most suitable as per the requirements
generated by the background. When this system falters, the next best available
alternative is used.”
Hence, it is to be noted once again that while various theories of Lexical Access
have been proposed, the factors involved in the process have been overlooked,
which are based on aspects such as Entrenchment – a Cognitive Linguistic notion,
among others as described above. Thus, this paper employs the Cognitive
Linguistics approach to carry out an overview of current theories, while also
suggesting the presence of a number of factors affecting the process. The next
chapter shall take into account each of these theories, as well as a study of how the
theoretically established factors such as Equivalence and Entrenchment function.
11
3 THEORETICAL BACKGROUND
3.1 COMMITMENTS TO AND OF THE APPROACH
In Cognitive Linguistics, language is considered a part of the human cognitive
structure, and the basic principle is to apply the same points to it that one would to other
cognitive processes. This is evident in the Cognitive Commitment and the
Generalisation Commitment as stated by Evans (2007), which are as follows:
-
Cognitive Commitment:
The principles of linguistic structure should reflect what is known about human
cognition from other disciplines, particularly the other cognitive sciences (philosophy,
psychology, artificial intelligence and neuroscience). It follows from the cognitive
commitment that language and linguistic organisation should reflect general cognitive
principles rather than cognitive principles that are specific to language. This
commitment, central to and definitional of cognitive linguistics, leads to the
generalisation commitment and the rejection by cognitive linguists of the modular
approach to language and the mind adopted in formal linguistics.
-
Generalisation Commitment:
It constitutes a commitment to the characterisation of general principles that are
responsible for all aspects of human language. This commitment follows from the
assumption central to cognitive linguistics that language reflects general cognitive
mechanisms and processes. Hence the generalisation commitment leads cognitive
linguists to search for common organising principles across different language
‘systems’, such as phonology, syntax, semantics and so on. Such common organising
principles include conceptual mechanisms like metaphor, conceptual blending and
phenomena such as polysemy. The generalisation commitment stands in direct
opposition to the modular approach taken in formal linguistics.
12
Given the above, then, it is important to note the difference between Cognitive
Linguistics and the schools of thought that are based on the Generative View
(Marantz, 2005) of the subject. The most important point of distinction is that while
Generative Linguistics views language as a separate modular system that works
simultaneously with other specific “modules” of cognition, Cognitive Linguistics
views the entire cognitive system as an inextricable whole, wherein various
processes occur across various areas. This is rather like the road networks in any one
of the vastly populated cities of the world where roads, lanes and highways are
connected in and across all directions, as compared to the streamlined lanes that one
sees on the race tracks, for instance. Basically, the idea is that for Cognitive
Linguists, the entire cognitive system is a holistic network, while for Generative
Linguists, it is a channelised, streamlined group of systems which are
simultaneously at play.
Thus, according to these guidelines, from the perspective of Cognitive
Linguistics as taken in this study, language is fundamentally a part of the cognitive
structure. Assuming this statement to be true, we now move on to the specificities of
this study.
In this paper, we study the way in which one specific part of Language
Production - Lexical Access, or in layman’s terms, word selection, occurs. How
does an individual select words with which to communicate? What factors affect
this selection process? What happens when this system malfunctions? How can
speech and usage-based errors be explained? All these questions create the building
blocks of the research potential of this study.
13
3.2 SUBSEQUENCE AND CASCADE/CONNECTIONIST MODELS OF LEXICAL ACCESS
3.2.1
Subsequence Model: Levelt and Schriefers (1987)
The most fundamental premise of this paper is the model presented by Levelt and
Schriefers (1987). Following in the authors’ footsteps, we consider the basic notion of
the “linguistic sign” as described by Saussure (Saussure, 1916); what the authors like to
call “Saussure’s Egg.” According to this theory, the linguistic sign was made up of a
“Concept,” as well as a Sound/Image pairing. Basically, it suggested a direct
relationship between encyclopaedic or “analogous” knowledge as later proposed by
Evans (Evans, 2009) and the phonetic as well as visual perceptual input referring to the
same. Essentially, then, the linguistic sign was considered to involve mainly a “formmeaning” pair (Saussure, 1916).
However, Levelt and Schriefers argue that this notion must also include the syntactic
or as Saussure puts it, “syntagmatic” properties of the sign in question, which they
define as “the ways in which the item can enter into phrasal combinations with other
linguistic signs” (Levelt & Schriefers, 1987). This indicates what syntactic properties
the word or group thereof might have, and perhaps even where it would occur within an
utterance. Thus, they suggest a flow-based model, wherein the Concept (Conceptual
Environment) interacts with the Syntax (Syntactic Environment) to create what they call
a “lemma,” which then interacts as a whole with the Sound/Image pairing (Phonetic
Environment, which they label the “lexeme.” According to Levelt and Schriefers,
The lexical item ‘resonates’ to the current conceptual environment, the speaker’s
speech act intention, message, or whatever it is called. When the item’s
conceptual conditions are sufficiently present in that environment, its syntactic
properties become available for the procedures of sentence generation. The
retrieval of other items is directly conditioned by the current syntactic
environment.
Hence, this theory suggests a scheme of things with input going from one component
to another, that is, from the conceptual module to the linguistic one, with the syntactic
properties following the semantic properties. However, as we have seen above, the
14
Cognitive Linguistics approach is quite different. Rather than viewing these components
as separate entities, we suggest they ought to be looked at as properties of the same
singular cognitive system. This is an important point in the proposed hypothesis, as the
proposition made here indicates the interaction of various factors in the process of
Lexical Access, which obviously includes the conceptual and linguistic aspects of the
word or phrase to be accessed.
Furthermore, the model suggested by Levelt and Schriefers describes the processes
that occur before and after the functioning of this modular system, which they call the
“lexical access proper,” stating that the activation of the lemma and lexeme could take
place either successively, or in a simultaneous manner. Given the approach described
above, however, the authors seem to lean more towards a distinction of the processes,
suggesting that the Conceptualiser is followed by the Grammatical Encoder, which in
turn is followed by the Sound Form Encoder, with the process terminating with the
Articulator. Later, Levelt suggested that the second and third stages, that is, the
Grammatical Encoder and the Sound Form Encoder must be melded together to form
the “Formulator” (Levelt, 1993). In this study, however, we do away with these separate
“informationally encapsulated” units that distinguish between phonological, syntactic
and semantic input. There are two main reasons for this. The first is our suggestion that
if language is based on and part of cognition and draws on embodied experience (as is
believed in the field of Cognitive Linguistics), it would follow that it also functions in a
similar manner. Hence, just as the latter receives perceptual and conceptual information
and upgrades on a constant basis which creates the foundation for their interaction with
their environment, language too would require a continuous stream of conceptual and
linguistic information in order to function correctly. The other reason is a corollary to
this suggestion – as we shall see in chapter 4, the process of Lexical Access is
somewhat cyclic in nature, involving input from a number of “dimensions” of language
at the same time, more than once, and in combination with each other, which would not
be the case if they were separated into informationally encapsulated units and not
allowed to interact at the same level. Articulation would possibly come later, although
there might be arguments regarding that, mentioning that there are many times when
people fumble for words halfway through the selected word, or even say the wrong one,
if it sounds similar, without realising the difference in speech, even if they knew the
difference in meaning.
15
Thus, while Levelt and Schriefer’s model seems to indicate a number of systems or
processes occurring in a subsequent manner to each other, A Cognitive Linguisticsbased approach would suggest a model that views these processes as simultaneous
dimensions of a single, holistic cognitive system. This shall be worked out in detail in
chapter 4.
Conceptualiser
•Maps
Communicative
Intention onto a
Preverbal Message
• Takes message as
input, produces surface
structure as output
Sound Form
Encoder
Formulator
Grammatical
Encoder
•Maps the surface structure onto
a phonetic/articulatory plane
Articulator
• Interprets and executes
the phonetic plan
Fig. 1: “Subsequence Model” – Levelt & Schriefers (1987), Levelt (1993)
3.2.2
Cascade Model: Dell, et al. (1999)
On the other hand, Dell, et al. (1999) describe a “cascade” model of Lexical Access,
wherein it is assumed that the process of “Phonological Encoding” begins before the
end of the seemingly preceding one of “Grammatical Encoding.” Also known as the
Connectionist Model, it indicates that semantic, syntactic and phonological information
is activated in a seemingly simultaneous manner, considering that a number of speech
errors, such as the Hungarian restaurant instead of Hungarian rhapsody instance,
involve phonological similarity within the correct and erroneous lexical items used,
despite the lack of a semantic connection between the two. The main claims made by
this model are:
i) That the two modules mentioned above do not work sequentially, but parallel to
each other.
ii) That the words chosen in the beginning of a sentence define the subsequent ones.
iii) That in contrast to comprehension, production models are required to “make
linguistic structure a priority.”
16
iv) That in case of speech errors, words that fall into the same “grammatical class” tend
to be substituted for one other, that is, nouns substitute nouns, and so on.
v) That the TOT (Tip-Of the-Tongue) phenomenon occurs after the stage of lemma
selection is over, and when phonological encoding begins; the speaker is unable to
retrieve the phonological units that constitute the lemma.
vi) That “lexical access is not a generative process,” and apart from “productive use of
morphology,” the required words are already part of one’s lexicon.
Conceptualiser
Grammatical Encoder
Sound Form Encoder
Articulator
Fig. 2: “Cascade Model” Dell, et al.. (1999)
With respect to the two models presented here, then, it can be surmised that while
each makes some valid and interesting arguments regarding the nature of Lexical
Access, which of course form the bases of our perspective here, there still remains the
fact that they both view language as a separate modular system, which clashes with the
tenets of Cognitive Linguistics. Also, there is the issue of the factors involved in the
process, which are not explored by either of these theories, and which we attempt to do
here. Hence, neither of the aforementioned models attempt to define precisely what
elements affect the process of Lexical Access, or talk about Semantic Specificity,
Context, Degree of Emphasis or Semantic/Articulatory Generalisation, or even points
such as Equivalence and Entrenchment. On the basis of these points, we question these
models and indicate that they should at least be considered incomplete, which leads us
to reject them as complete theories.
In order to work out the functioning of Lexical Access from the conceptual system to
the linguistic one, and given the relative lack of resources to draw upon from the
Cognitive Linguistics point of view, we first need to select and assess one of the most
popular recent theories of conceptual access through lexical items, and then see if they
could be applied to the process of production, when turned ‘upside-down,’ so to speak.
17
One theory stands out the most in terms of Access, albeit from the comprehension
point of view. This is the LCCM Theory as laid out by Evans (2009), which we shall
now discuss in the next section.
3.3 THE LCCM THEORY
The Theory of Lexical Concepts and Cognitive Models (also known as Access
Semantics), as explained by Evans (2009), is one describing the processes that
seemingly occur within the cognitive system during the comprehension of language,
while addressing mainly the “protean nature of word meaning.” The theory itself is
comprehensive and exhaustive in nature, taking into account all apparent aspects of
cognition within the comprehension process. We shall first take a look at the claims
made by Evans within this premise, and then move on to working out exactly how it
could be employed for the process of production as well.
We begin with the overall structure of Access Semantics.
Evans suggests an approach towards “accounting for the inherent variation in word
meaning” through “a principled separation between the linguistic system – the
linguistic knowledge that words encode – and the conceptual system – the non-linguistic
knowledge that words facilitate access to” (Evans, 2009, p. xi).
He also indicates that while there are many theories attempting to connect semantic
and conceptual structure and join them both within the same group, these two aspects
are quite distinct from each other, with the former forming the linguistic content which
naturally is language-dependent, and the latter, the conceptual content which is
language-independent. However, he explains through logical argument, that conceptual
content cannot be context-independent as such.
Below is a summary of the Theory of Access Semantics, as put forth by XXX (2016)
•
Lexical representation is made up of Symbolic Units and Cognitive Models;
•
Symbolic Units consist of Lexical Concepts and Vehicles (which are
individual linguistic units);
18
•
Lexical Concepts refer to the Semantic Structure, which is parametric in
nature. This means that their content, independent of conceptual
connections, is highly schematic knowledge that is abstracted from
Cognitive Models;
•
Cognitive Models refer to the Conceptual Structure, which is the
encyclopaedic or, as Evans prefers to call it, “analogous” knowledge that is
present within the conceptual system, which is based on embodied
experience and is reactivated or “simulated” when we think;
•
The Access Site refers to a Cognitive Model that “facilitates direct access
from an open-class Lexical Concept to its relevant Cognitive Models;”
•
A Lexical Profile is the unique profile of each Lexical Concept, involving
the Use Potential;
•
A Cognitive Model Profile is the Semantic Potential – the set of Cognitive
Models that get activated based on the linguistic and non-linguistic context;
and
•
There are specific Semantic and Formal Selectional Tendencies that govern
the overall process of Lexical Concept Selection, which is followed by
Fusion, which in turn involves Lexical Concept Integration and
Interpretation.
Thus, according to the LCCM theory, there are two main nodes in terms of
“representation” when it comes to the interaction between language and the conceptual
framework within the realm of cognition. The first node, which Evans calls “Lexical
Representation,” is made of “Symbolic Units” and “Cognitive Models.” He describes
three basic principles of Lexical Representation (Evans, 2009, P. 30):
•
Polysemy is conceptual in nature
•
Words are associated with selectional tendencies (Lexical Profile)
19
•
Grammatical categories have a semantic basis
According to Evans, meaning is distinct from Lexical Representation as “while
meaning is a property of an utterance, lexical representations consist of the mental
abstractions which we infer must be stored as part of the language user’s knowledge of
language: symbolic units, together with the range of cognitive models, the semantic
potential, to which a lexical concept affords access” (Evans, 2009, p. 73).
Essentially, he explains that Lexical Representation consists of a more complex web
of aspects that create the bridge between the conceptual and linguistic systems, thereby
creating meaning for an utterance, or as Evans puts it, a “conception.” It is also
mentioned that “backstage processes” such as conceptualisation lead to meaning
construction. Evans calls this process of conceptualisation “simulation,” wherein a word
or utterance triggers conceptual content that is based on embodied experience. Let us
take the example of the word left, in the context of embodied direction, so to speak. The
Lexical Concept for this term, then, provides access to the cognitive models associated
with the aspects of “directedness” (Evans, 2013), and in doing so, gives rise to an
“image schema” (Johnson, 2005) of directionality in terms of ‘right’ or ‘left’ on the
basis of personal individual experience. An interesting thing to note here is the
connection between cognitive models associated mainly with activities that are specific
to the right or left hand, depending also on the cultural context, and their effect on
comprehension. In the Indian context, the [Hand/Eat] parameter of the [DIRECTION]
Lexical Concept would be associated with the right hand, whereas in most Western
cultures, which employ the use of cutlery as compared to Indians’ use of their fingers, it
would be associated with the left hand. Hence, the association of specific parameters of
Lexical Concepts with a Cognitive Model Profile is not only based on embodied
experience, but also culture-specific, convention-driven behaviour. An amusing factual
instance comes to mind; an Indian individual who usually has trouble telling left from
right once ended up sending people asking for directions the wrong way because they
forgot to carry out the ‘eating motion’ with their hands, to figure out which was which,
before giving the directions. This shows how strongly these Cognitive Models can
affect the overall process.
•
Aspects of Lexical Representation:
20
i) Symbolic Units: This term refers to a system that involves “Vehicles” and
“Lexical Concepts.” Essentially, a symbolic unit is a merger of the schematic
information that makes up the linguistic item, and the phonological form it
takes, eventually adding up to create the linguistic essence of the item.
a. Lexical Concepts: These include the schematic and “digitised” content
which is “abstracted from usage events” (Evans, 2009, p. 128). The
knowledge represented here, then, is purely linguistic and “parametric” in
nature. This takes into account whether the item needs to represent a
THING, PROCESS, RELATION (Taylor, 2002) or any other such
phenomena, which in the most basic understanding of grammar may be
considered as whether the lexical item belongs to the noun, verb, adjective,
adverb or other such lexical class. However, it must be noted that Lexical
Concepts themselves, when employed in usage, are only partially
activated; in fact, the context and usage-based requirements would
prescribe which form and linguistic aspects of the Lexical Concept would
come to the forefront. Evans (2009) likens this to the distinction between
phonemes and allophones, wherein phonemes are abstract entities with
specific usage-based options for realisation, and allophones are those
actual realisations in context. An interesting case is the lexical item walk,
which can be considered a THING (noun) or a PROCESS (verb),
depending on the utterance context. The question that arises here is
whether it would appear as the same meaningful abstract entity (Lexical
Concept) [WALK], with specific aspects of its linguistic contextual
parameters being activated, or if it would occur as two different Lexical
Concepts, each defined by its own linguistic parameters and thus
distinguished from the other, despite being drawn from within similar (or
congruent – it would be a matter of perspective) Cognitive Model Profiles.
Evans seems to suggest the latter, which shall be discussed later. Thus,
Evans proposes that Lexical Concepts constitute the “semantic pole” of
Symbolic Units. He also suggests that Lexical Concepts have their own
unique “Lexical Profile,” which consists of their “Use Potential” in terms
of the linguistic context.
21
b. Vehicles: Independent linguistic units that are employed to represent
lexical concepts and their related analogue content are known as vehicles.
This involves the phonologically bound entities that are packed together to
form morphemes, words or idioms – any linguistic structure that can be
viewed as a whole entity. Hence, Vehicles are known as the “linguistic
pole” of the Symbolic Unit.
ii) Cognitive Models: This aspect of Lexical Representation deals with the
“analogous” or “encyclopaedic” knowledge, which is non-linguistic in nature
and is specific to the conceptual system. Basically, it refers to the set of
conceptual structures that can be activated with the usage of a certain lexical
concept, depending on the context thereof. According to Evans, the use of a
particular lexical concept in a specific context gives rise to a number of
conceptual structures, which he calls “cognitive models,” which are clusters of
non-linguistic, encyclopaedic knowledge that we gain through embodied
experience, and which are activated through the means of an “access site,”
based on the linguistic and non-linguistic context. The entirety of the Cognitive
Models that are associated with a Lexical Concept regardless of context, is
called a “Cognitive Model Profile” for that particular Lexical Concept. This
forms the “Semantic Potential” for the same.
•
Aspects of Semantic Composition:
The second node that forms the LCCM Theory (Evans, 2009) involves the
processes that interact with the elements within the Lexical Representation node,
and is called Semantic Composition. This refers to “the linguistically mediated
meaning construction process” (Evans, 2009, p.75). It includes two basic
processes – Lexical Concept Selection and Fusion.
22
According to the theory, “a vehicle may potentially be associated with a
large number of distinct Lexical Concepts” (Evans, 2009, p. 217). This property
of each vehicle - being associated with an array of Lexical Concepts that exist
parallel to each other, is known as the Lexical Concept Potential of that
particular vehicle. Hence, it is indicated that this Lexical Concept Potential must
be narrowed down to a single Lexical Concept in order to produce a “reading,”
which Evans describes as “a situated interpretation specific to the context in
which it is embedded” (Evans, 2009, p. 218). Basically, it refers to the specific
conceptual structure(s) that would be associated with a vehicle within a given
context during comprehension; the latter would include the other Lexical
Concepts involved within the utterance.
Thus, Evans claims that in order to achieve this objective of reading, the
Lexical Concept Potential of a vehicle must undergo a kind of filtration, which
includes the processes of Lexical Concept Selection (mostly referred to as
“Selection”) and Fusion, the latter of which includes two sub-processes, namely
Integration and Interpretation. These sub-processes seem to be governed by a set
of 10 basic principles in all, which shall be discussed later in the chapter.
i) Lexical Concept Selection: This process, as is evident from the term
itself, involves the identification of a specific Lexical Concept for a
specific Vehicle in a specific context, within a given utterance. It is
suggested that Selection occurs between distinct Lexical Concepts and
not at the level of the Semantic Potential (Cognitive Models). Essentially,
Selection occurs when there is more than one Lexical Concept available
for a vehicle, each with distinct Lexical Profiles and consequently a
number of potential Access Sites, as compared to ones with a single
Lexical Profile and similarly, a single Access Site. Let us take the
example of some adjectives/adverbs that are related to visual perception.
i. Shiny:
a. That’s one shiny diamond!
b. She gave him a shiny smile.
23
In the sentences given above, the Lexical Concept of [VISUALLY
BRIGHT/ LUSTROUS] seems to be at play in the first instance, whereas
a more metaphorical usage is indicated in the second one, perhaps with
the Lexical Concept [SEEMINGLY HAPPY] in action. While this
suggests the existence of more than one Lexical Concept, Evans’
explanation seems to cover only the broad range of parameters used to
define and distinguish one Lexical Concept from another. This point is
discussed more elaborately at the end of the chapter. The process of
distinguishing Lexical Concepts may cause a certain degree of confusion
while deciding how many Lexical Concepts a vehicle can have,
especially when it comes to homonymy or polysemy. While Evans does
address the issue of polysemy in quite a lot of detail, the focus remains
on the fact that different Lexical Concepts may well exist for a word, and
not on how they are defined. He also mentions, as stated above (Evans,
2009, pg. 30), that polysemy is “conceptual in nature.” However, this
gives rise to the question of how exactly Lexical Concepts are defined for
words that have distinct, but related meanings.
ii. Dull:
a. The silver looks dull and dusty.
b. This is such a dull job.
In the first utterance, dull seems to mean [OLD/VISUALLY
UNAPPEALING], whereas in the second, it suggests the use of the
[BORING/MUNDANE] Lexical Concept.
Given these examples, we can now work out the overall process of
Lexical Concept Selection – a vehicle, when uttered, generates a number
of possibilities in terms of Lexical Concepts within the hearer’s mind, out
of which one (or more) are selected, giving rise to the Cognitive Model
Profile of that particular Lexical Concept or blend thereof.
The first phase, then, Evans calls “broad selection,” wherein one
Lexical Concept is selected from an array of potential ones. Next comes
24
the process of “narrow selection,” wherein he suggests that each Lexical
Concept has a number of different “parameters,” which would define
specific usages and inferences, and from which one (or more) parameters
would be engaged and selected. Consider the following example.
iii. Tall:
a. He is tall for his age.
b. The city is full of tall buildings.
c. She tells tall tales.
In this scenario, it seems that the vehicle tall activates the selection of
a single Lexical Concept in each of these utterances – [OF GREAT(ER)
HEIGHT (THAN AVERAGE)]. Even on a figurative level – the “tall
tales” seem to refer to the same Lexical Concept, except that it seems to
focus on the part about the tales being ‘beyond the reach of belief,’ and
thus unrealistic or not believable. Hence, the Lexical Concept [OF
GREAT(ER) HEIGHT (THAN AVERAGE)] would have at least two
parameters associated with it – that of visual perception and that of
something being metaphorically out of reach – both of which are
connected in the fact that something or someone tall would generally be
beyond the reach or potential of the similar things or people surrounding
it. There may be a certain amount of debate over whether this explanation
is acceptable or not, depending on the approach taken. A similar situation
occurs in the case of short, as seen below, where the Lexical Concept
involved is [OF LESS(ER) HEIGHT/LENGTH (THAN AVERAGE)]. A
point to be noted here, then, is that while tall and short are essentially
considered antonyms (Turney, 2008) or opposites of each other, the latter
seems to also include the aspect of (usually horizontal) ‘length,’ to add to
the vertical ‘height aspect.
iv. Short:
a. He is short for his age.
b. It seems he fell short of his parents’ expectations.
c. Keep it short.
25
-
Context: There are a number of factors that affect Lexical Concept
Selection, which are divided into two broad categories: Linguistic
Context and Extra-linguistic Context. The former includes the Utterance
Context, which denotes the linguistic and prosodic features involved in
the utterance, such as intonation, rhythm and stress etc. It also includes
the Discourse Context, that is, the “common ground” or shared
knowledge between the speaker and hearer(s), which is textual and
situational in nature, and the Speech Event, which leads to selection
based on the situation in general. The Extra-linguistic Context would
involve the overall venue, setting etc. of the utterance. For instance,
while Evans uses the example of expired to explain this phenomenon
(Evans, 2009, p. 221), we make use of the polysemous word goal –
which can mean different things in different situations. We consider two
different scenarios here – a football game and a meeting of an
organization’s top executives. In the first scenario, the following would
be considered in terms of Context:
(i) Football Game:
1. Extra-linguistic Context: The football game taking place in a
stadium with hundreds of fans for both teams, cheering them
on.
2. Linguistic Context:
a. Utterance Context: The intonation and stress patterns of
the utterance would lead to the construing of an
exclamation on the audience’s part.
b. Discourse Context: The common ground would be the
knowledge shared by football fans regarding the rules and
terminology of the game, thus leading them to know that a
goal in this context would involve one of the team
members sending the ball through the post and into the
net, and past the goalkeeper.
26
c. Speech Event: Again, considering the overall situation,
vis-à-vis the commentary during a football game, the
“meaning” of the word goal becomes quite clear.
(ii) In the boardroom meeting situation, the inference drawn would
be based on the following.
1.
Extra-linguistic Context: A small number of people sitting in
a conference room, with one individual (perhaps the CEO, or
someone with that kind of authority) speaking to the rest of
them regarding the company’s functioning and vision.
2. Linguistic Context:
a. Utterance Context: The intonation, rhythm and stress
involved in this utterance would suggest a declaration,
possibly said with an air of authority.
b. Discourse Context: The common ground would be the
shared knowledge of the various aspects involved in the
functioning of a company, and the fact that there are
targets and objectives to be met.
c. Speech Event: This would involve the talk given by the
individual, perhaps with the motive of offering inspiration
or reprimanding negligence.
In either situation, then, it would be rather difficult for people to mix up
what is meant by the word goal, despite it being the same sound form.
We now move on to the next process of Lexical Access – Fusion.
ii) Fusion: This section involves two basic processes – Lexical Concept
Integration and Interpretation. Lexical Concept Integration as a process
forms composite units known as Lexical Conceptual Units – integrated
units of linguistic content. Essentially, Evans suggests that it provides the
“scaffolding” for the activation of conceptual content. He distinguishes
between two types of open-class Lexical Concepts – Internally Open
(abstract) and Internally Closed (concrete). According to this view, the
27
former kind are not “lexically filled,” that is, they do not have “fixed
lexical items as part of their composition” (Evans, 2009), whereas the
latter do. He also suggests that in case of the former kind of Lexical
Concept, there occurs the process of Internal Lexical Concept
Integration, while the latter involves External Lexical Concept
Integration. Both these processes create a holistic Lexical Concept for the
utterance itself, after going through individual integrative processes for
each separate Lexical Concept within the utterance. This is a recursive
process, wherein once the Cognitive Models for each Lexical Concept
have been activated, they are put together with the others within the
utterance, and if there are no clashes among them in terms of conceptual
or linguistic structure, then the entire utterance is considered as a whole
and interpreted accordingly.
There are 10 basic principles (Evans, 2009) that govern the process of
Fusion. These are as follows:
•
Principle of Linguistic Coherence: “A lexical concept that is internally
open may only be integrated with a lexical concept with which it shares
schematic coherence in terms of linguistic content.”
•
Principle of Schematic Content: “The content associated with entities,
participants, and the relations holding between them must exhibit
coherence in fusion operations.”
•
Principle of Ordered Integration in Internally Open Lexical Concepts:
Lexical concept integration takes place by applying to internally simpler
lexical concepts before applying to internally more complex lexical
concepts.”
•
Principle of Guided Matching: “Matching of cognitive models in
interpretation proceeds in a way that is compatible with the output of
lexical concept integration.”
•
Principle of Conceptual Coherence: “Matching occurs between one or
more cognitive models/informational characterisations, belonging to
28
distinct cognitive model profiles/lexical conceptual units, which share
schematic coherence in terms of conceptual content.”
•
Principle of Schematic Salience in Matching: “Matching across cognitive
model profiles/informational characterisations achieves greater schematic
salience when relatively more cognitive models are matched than
matches involving fewer cognitive models.”
•
Principle of Simultaneous Matching: When matching takes place
between an informational characterisation and a complex lexical concept,
matching may occur simultaneously across cognitive model profiles of
the lexical concepts that form a part of the complex lexical concept.”
•
Principle of Primary Activation: Matched cognitive model(s) are subject
to primary activation.”
•
Principle of Ordered Search: “Matching takes place in the primary
cognitive model profile, which is the default search region for that subset
of lexical concepts that facilitate access to a cognitive model profile. If
matching is unsuccessful in the default search domain… a clash occurs,
then a new search domain is established in the secondary cognitive model
profile. The search proceeds in an ordered fashion, proceeding on the
basis of secondary cognitive models that are conceptually more coherent
with respect to the primary cognitive models (and hence modelled as
being ‘closer’ in the cognitive model profile) prior to searching cognitive
models that exhibit successively less conceptual coherence,”
•
Principle of Secondary Activation: “All primary cognitive models, and
all secondary cognitive models on the route of access which do not
achieve primary activation, achieve secondary activation.”
There still remain some points to consider when it comes to the LCCM Theory.
First of all, there is the claim that closed class elements are less content-based than
open class ones. For instance, we take the utterance The men are going to war. In this
sentence, if the open class elements are removed, we get “The somethings are
something-ing to something.” This suggests that a specific group of individuals is
currently carrying out an action that culminates at a specific point. The same
structure could be used for The girls are heading to school, or The boys are talking to
Dexter, etc. However, the information we gain from the closed class elements is still
29
conceptually relevant. We make the argument that words such as the, are and to are
equally conceptually-driven as the others in any of these sentences. True, they may
not have an entire array of Lexical Concepts at their disposal, but they would still be
related to the cognitive system, in terms of providing more than just schematic
information. For instance, the leads to a good deal of conceptual information, which
is drawn from the cognitive models that define the existence of specificity, which
leads to the definiteness of the article, as well as what it describes – in this case, men,
girls or boys. Just like the ‘meaning’ of the utterance would change if the open class
elements were changed, it would do the same for closed class ones too, which is seen
if we replace the with these or those. There may be a smaller degree of difference,
but it would still exist. Aspects such as definiteness, specificity, tense, number, and
prepositional value are all vital in terms of meaning, and ought to derive their content
from cognitive models, as they are not purely linguistic or schematic in nature, as is
suggested. The same goes if we replace to with for here, especially in the third
utterance. Also, an argument could be made regarding the limited number of
alternatives that could be chosen for closed class elements as compared to open class
ones. However, given the structure of any utterance, the existence of closed class
elements tends to constrain the alternatives for the open class elements as well. For
instance, while the slots given above may be filled with girls, heading and school or
boys, talking and Dexter, they cannot be filled with random words falling within the
same grammatical category, such as cats, creeping and wall, without adding another
the before the last word. This suggests that closed class elements are also
conceptually-driven, as are open class ones. The biggest examples of this are seen in
Evans’ use of in, on and over as Lexical Concepts involving polysemy.
Another point to note, as mentioned above, is the fact that while more than one
Lexical Concept may well exist for a single vehicle, it is important to define the
characteristics of Lexical Concepts themselves, outlining the parameters that govern
the distinction of one from the other. Evans (2009, p. 141-142) describes two basic
distinguishing criteria for Lexical Profile selection, and in turn, Lexical Concept
selection:
-
The Semantic Selectional Criterion:
A distinct lexical profile – by definition encoded by a distinct lexical concept –
provides unique or highly distinct patterns in terms of the nature and range of
30
the lexical concepts with which a lexical concept can co-occur or in which it can
be embedded, or in the case of an internally open lexical concept, which occur
within it.
-
The Formal Selectional Criterion:
A distinct lexical profile – by definition encoded by a distinct lexical concept –
provides unique or highly distinct patterns in terms of the vehicles with which a
lexical concept can co-occur or in which it can be embedded, or in the case of an
internally open lexical concept, the nature of the alignment between vehicles and
the internally closed lexical concepts that lexically fill the internally open lexical
concept.
However, despite these criteria, sometimes differentiating between Lexical
Concepts can be difficult. An instance of this is seen in Evans’ (2009, p. 220-221)
own example of the Lexical Concepts associated with fast.
(1) a fast car [RAPID LOCOMOTION]
(2) a fast typist [RAPID PERFORMANCE OF ACTIVITY]
(3) a fast decision [REQUIRES LITTLE TIME FOR COMPLETION]
(4) …a fast garage… [Blend: all of the above]
While all of these are considered distinct Lexical Concepts for the word fast in the
given description, if it comes to the Semantic Selectional Criterion, the basis of
Lexical Concept Selection here would be the slight differences in semantic value.
However, Evans also rightly points out that if every usage of the vehicle is
considered as a distinct Lexical Concept, the list would be infinite. The point to note
here, though, is that while one can understand why the Lexical Concept is different
for (3), given the use of the Time domain, (1) and (2) are based mainly on
Motion/Movement. Whether it is locomotion or the performance of an activity, the
basis of both potential Lexical Concepts is [MOTION], and thus makes one wonder
why they have been separated in the first place.
Also, considering that the Lexical Concept is supposed to be highly schematic and
linguistic in nature, the conceptual nature of its collocations should not really be a
point of concern. Thus, while the Lexical Concepts for car and typist differ only in
that the former defines an inanimate THING and the latter defines an animate one,
their respective Semantic Potentials should not be relevant. The Semantic Selectional
Criterion mentions the uniqueness of patterns in terms of “nature,” which makes one
31
wonder what exactly constitutes “nature” here. If it is schematic content, then the
only difference in the Lexical Concepts for car and typist is that one is inanimate,
while another is not.
However, even this knowledge is to a degree conceptual in nature, based on the
embodied experience and analogue information that a car is distinct from a person,
and while both move, the former requires another entity to bring it into motion, while
the latter does not. Thus, from the purely schematic perspective, we are only looking
at [THING] here, perhaps with certain other aspects such as number, tense etc. This
seems to entail that –
a. Lexical Concepts are not purely schematic in nature; they also include
conceptual content. If that is the case, however, then the use for fast or any other
adjective to describe each and every [THING] in its collocations would incite the use
of a distinct Lexical Concept – making the list infinite, or
b. Lexical Concepts are truly schematic in nature, which suggests that the
Semantic Potential of their collocations or any part thereof must not interfere with
the selectional criteria, and would thus suggest a single Lexical Concept for common
use, unless there is a notable difference in its own semantic value, as seen in
homonomy.
We suggest a third alternative, at least in terms of Production, as we shall see in
the next chapter – taking aspects from both options given above. If the Lexical
Concept itself is considered a part-schematic, part-conceptual category for a range of
potential vehicles to be used within different usage-based contexts, then both these
problems can be solved. Hence, the schematic information encoded by it would take
first preference, creating a category of options for the vehicles describing (in case of
an adjective) a RELATION (property) to/of a THING. Having established that, the
Semantic Potential of the Lexical Concept would come into play, eliminating various
vehicles that do not fall within the required semantic range.
Hence, if placed under the Lexical Concept category of [RAPID], the issue could
be sorted out more easily than the divisions given above, with the collocational
Lexical Concept being [MOTION]. If it was about thinking fast, that too is
essentially an “activity,” but focuses more on the aspects of Time and Cognition, and
hence, would possibly be put together with (3) under the same context. In terms of
32
comprehension, then, the vehicles with more than one truly distinct Lexical Concepts
would select, based on their semantic value, the one best suited for the context, and
access the combined Semantic Potential that is generated by the connection between
the two accordingly. While this is pure conjecture and exploring it is beyond the
scope of this paper, further studies carried out in this area would certainly yield
interesting results.
Given below is a pictorial representation of Evans’ LCCM Theory as described by
XXXX (2016) –
LCCM
Theory
Lexical Representation
Cognitive Models
Symbolic
Units
Vehicles
Lexical
Concepts
Selection
Fusion
Semantic Composition
Integration
Lexical Conceptual
Units (Semantic
Value)
Interpretation
Fig. 3: The LCCM Theory
Access Site
Cognitive Model
Profiles
33
3.4 SYNONYMY
The concept of synonymy has been a long-standing matter of debate among
linguists, whatever their approach. While some early linguists believed that the
semantic and formal values of one word can be “similar” to those of another, others
later pointed out various issues related to the claim (Martin, 1984). For one thing, it
is suggested that seemingly synonymous words may have related meanings, but are
distinct in terms of some conceptual and semantic nuances. For instance, if we take
the most commonly occurring adjective in the British National Corpus (Sketch
Engine, 2004) – other, the top 3 “synonyms” we get from the Sketch Engine
thesaurus are different, another/some other and new – all of which differ to a certain
degree in terms of actual conceptual and “semantic value.” The term “semantic
value” here can be defined as the “exact meaning” to be delivered by the word,
which involves a specific set of conceptual structures that are based on embodied
cognition and experience, and which make up a unique combination of semantic
characteristics associated with it. Similarly, if a cross-linguistic analysis were to take
place here, the “synonyms” for [ənjə] (which loosely translates into other in Hindi)
would be [du:sra] (second/another), [ələg] (different) and [nəja] (new) or even
[bɦinn] (distinct) (Sketch Engine, 2004), if considering the formal register of Hindi.
From inherent knowledge of the languages (as a native speaker), it can be surmised
that the same would occur in Gujarati and Marathi, leading to [ʤudũ] and [ʋegɭə]
being
considered
“synonymous”
to
[biʤũ]
(second/another)
and
[həʤi]
(more/someone or something else), as well as [dusrə] (second/another), [əʣu:n]
(more) and [aɳkhi] (further/more).
From the point of view of the LCCM Theory, then, it can be considered that
synonymy per se doesn’t exist. While there may be a number of overlapping
cognitive models between two or more words or phrases (not necessarily those
within the same grammatical category either), each Lexical Concept would have its
own Cognitive Model Profile, and thus, even if we discount the “Use Potential” of
the Lexical Concept by arguing that there are various scenarios in usage wherein
despite the use of a different word or phrase the content of the utterance and the
speaker intent are conveyed successfully, the “Semantic Potential” thereof will still
be distinct from others falling within a similar ‘semantic range,’ so to speak.
34
‘Semantic range’ here refers to the array of conceptually associated semantic values
for various lexical items, which range between least to most similar, depending on
the context.
While all these points are difficult to refute, and seem to make sense, there also
exists the fact that eventually, the aim of language is to act as an “instrument for
organising, processing, and conveying information” (Sanders and Spooren, 2007, p.
916). This brings forth two main arguments. The first is that while the Semantic
Potential matters at the speaker level, this potential for the same word or phrase may
be slightly different for the listener, and there cannot be complete congruence in
these two, given the statement above, which indicates that conceptual structure is
based on embodied experience. Since no two human beings can have the exact same
embodied experiences throughout their lives, complete congruence is not possible.
This means that to a degree, the listener is adding on to his/her own understanding of
the word or phrase in question by attaching a new ‘meaning’ or context to its use and
possibly even giving rise to polysemy through such learning. On the other hand,
there is also the point that at the utterance level, the context seems to be the most
important aspect. The use of any word or phrase from within the same semantic
range, given certain basic constraints and the extra-linguistic factors mentioned
above, can be permissible as long as their linguistic context is clear, and fits with the
extra-linguistic context.
A good example of this is the word thing or in more recent use, even thingy, to
describe a physical or abstract object or entity. For instance, take the statements –
i.
Hand me the broom, as opposed to
ii. Hand me the thing.
Within the same linguistic (prosodic) context and with the use of the same
gestures within the same speech situation (Evans, 2009), thing can still be used to
substitute broom in order to convey the relevant message. On the other hand, in case
this doesn’t work and the hearer does not understand which thing” the speaker is
talking about, it incites the former to interact further, asking for a more specific
35
description. In case the speaker is experiencing a Tip-of the-Tongue (TOT) (Brown
& McNeill, 1966) phenomenon at this time, they may just describe the broom as
iii. The thing you sweep with.
This would lead to better understanding on the hearer’s part. This is where the
aspect of categorization also comes in. It can be argued that a broom still falls in the
category of a thing, and thus the hearer has at least some linguistic or as Evans puts
it, “parametric” knowledge of the seemingly synonymous words, and their potential
relation to each other. It also indicates that there is a specific Cognitive Model Profile
at play, which makes use of words that define the function of the object being
discussed, and thus provides at least a somewhat detailed description of it. However,
it may also be argued that a number of objects can exist together with the broom, in
the same general direction that the speaker gestures towards, and unless there are
visual or other such cues suggesting precisely which thing the speaker really needs, it
becomes an ambiguous statement that demands further clarity. For the most part,
though, given the extra-linguistic context, it is usually possible for successful
communication to occur.
The notion of “cognitive economy” is also relevant here (Evans, 2007). It is
suggested in the field of Cognitive Linguistics that while there is a great deal of rich
conceptual content within a speaker and/or hearer’s mind, the speaker would usually
try to express themselves using as specific a conceptual structure and categorisation
tools as possible, which in terms of the LCCM Theory, would mean a very specific
part of the Cognitive Model Profile and thus, the Lexical Concept related to it, in
order to convey the message. However, in phenomena such as the TOT situation, the
same conceptual structure may have to be broadened to include other aspects of the
Cognitive Model Profile and enable “substitution” by apparent “synonyms” or other
such equivalents (see next section). This could lead to further Generalisation of the
overall content thereof, and consequently, necessitate the choice of words expressing
other characteristic features of the object/entity in question in order for the speaker to
express themselves more efficiently. Continuing the example of the broom here, it
36
can be said that in case the TOT phenomenon occurs during the speaker’s utterance,
they might first use a more generic term to substitute the word broom, such as thing.
If due to contextual “noise” (interference by entities or phenomena within the context
that carry a similar description, function or other defining value) this does not work,
it would then be focused on a different part of the cognitive model profile (in this
case its function) involving other aspects. It will thus result in the selection of the
words most specific to expressing this new functional description. This means that
the shift in focus within the Cognitive Model Profile would lead to the selection of
words expressing other aspects of the object/ entity in question.
3.5 EQUIVALENCE
As a related concept to synonymy, Jakobson’s (1959) theory of equivalence seems
to pave the way for a more comprehensive understanding of the phenomenon given
above. While his claims were made on the basis of the study of Translation, the
points he has stated are the building blocks for the arguments laid out in this study.
The most important aspect thereof is that of “intralingual translation.” Jakobson
enumerates three basic kinds of translation – “intralingual,” “interlingual” and
“intersemiotic.”
• Intralingual Translation: This kind of translation, also termed “rewording,”
occurs within the language, and can be described as the substitution of a word or
phrase with an apparently “synonymous” one. This could also occur in case the
target word or phrase is for any reason unavailable, as in Tip-of-Tongue situations.
The author defines it as “…an interpretation of verbal signs by means of other signs
of the same language.” Essentially, as is currently happening to a degree within this
very sentence, it refers to what is generally known as “paraphrasing.” This concept is
one of the most vital cogs of the Lexical Access system, as it provides the basis for
the substitution of specific words within the lexicon with others that have a similar
semantic value. It follows, then, that the Cognitive Model Profiles associated with
this type of translation are either slightly different, but overlap to quite an extent, or
as mentioned in the previous section, simply require a shift in “focus” to other
aspects within them. In case of word-for-word substitution, this would involve the
37
use of Lexical Concepts that are associated with each other through their respective
Cognitive Models, and thus, even if the substitute word does not sound exactly “apt”
for the context or changes the meaning ever so slightly, it can still be used, as it
conveys the basic message. As mentioned above, it is the linguistic and extralinguistic context that helps the hearer understand the semantic value of the target
word and apply it to the Cognitive Models within their own conceptual structure.
• Interlingual Translation: This kind of translation refers to the substitution of
words or phrases within one language with those in another. Jakobson calls this
phenomenon “translation proper,” indicating that it is “…an interpretation of verbal
signs by means of some other language.”
• Intersemiotic Translation: Here, translation occurs between two distinct semiotic
systems, that is, from verbal to nonverbal systems, in order to convey meaning. An
instance of this is the adaptation of prose or poetry into dance, wherein every word or
phrase is interpreted in terms of gestures and/or movements. Another example is the
existence of Sign Language, as we see in a number of news channels where an
individual gesticulates while the reader reads out the script. The author describes this
as “transmutation,” suggesting that it is “…an interpretation of verbal signs by
means of signs of nonverbal sign systems.”
While the first two kinds of translation are relevant to Lexical Access in Language
Production and describe the concept of “equivalence” quite well in order to define
the equality of semantic value between two distinct words or phrases, the correlation
between them both is a matter beyond the scope of this paper. Thus, we shall only
consider the point of Intralingual Translation for this specific study, as it focuses on a
more in-depth analysis of synonym substitution as compared to the others.
3.6 EQUIVALENCE FROM A COGNITIVE LINGUISTICS PERSPECTIVE
Carrying on the point of Intralingual Translation, then, we come to a bit of an
impasse as to the level of “loss” that occurs during the process, with various theories
suggesting various approaches to it (Halverson, 1997). Here the term “loss” refers to
the loss of semantic value when one word or phrase is substituted by another.
However, while this debate is by no means inconsequential, there exists another
38
point to all of this – that of ‘gain’ in semantic potential. As such, with the suggestion
that different vehicles are connected to distinct Lexical Concepts which provide
certain conceptual structures associated with them, it follows that while there would
certinly be a loss of certain Cognitive Models from within a profile, there would also
be an addition of other Cognitive Models into the revised profile, when it comes to
the use of a substitute. The other option is that only the vehicle is mapped onto the
Cognitive Model Profile of a different word or phrase, leading to a complete loss (for
the substituted vehicle in terms of semantic potential) and no gain in terms of
conceptual structures in the speaker’s mind. This seems rather far-fetched, though. If
a vehicle is inextricably connected to a Lexical Concept, then it follows that the
Cognitive Model Profile would tag along when it is used in an unfamiliar context and
would evolve accordingly, given that it is because of the Cognitive Model Profile
that the substitute word or phrase was selected in the first place.
While we have spoken mainly about the gain in semantic potential in terms of
word-for-word substitution here, it is also important to note that in carrying out the
process, there is a certain degree of ‘linguistic gain’ occurring for both, the speaker
and the hearer’s conceptual systems. This means that despite the seemingly
inappropriate usage of the substitute word, it could be added as an instance of how
the relevant conceptual structure could be expressed, and if repeated a number of
times, the vehicle and its respective Lexical Concept would become part of their
conceptual/linguistic repertoire (see next section).
In this study, we carry on the distinction between what is described as “formal
equivalence,” and “dynamic equivalence” by Nida (2003). He describes the former
as focusing on both, the form and content of the message, and the latter as a
phenomenon in which the focus is on the relationship between the hearer and the
message, which should essentially be the same as was with their original counterparts
(Nida, 1964). Essentially, he distinguishes between the kind of equivalence that deals
with both, form and content, and the kind that focuses entirely on content, as long as
the message is communicated effectively – that is, if it has the same comprehension
value as the original message in the hearer’s mind. When we say “value” here,
though, it does not necessarily mean that it activates the exact same Cognitive Model
Profile as that of the original word or phrase, but it is semantically as close to it as
39
possible, allowing the linguistic and extra-linguistic context to fill any gaps left by
the substitution.
While this dichotomy between formal and dynamic equivalence may be
conducive from the modular approach, given everything discussed above, it seems
plausible that a greater degree of focus is set on content when it comes to
substitution, rather than form.
3.7 ENTRENCHMENT, COLLOCATIONS AND SALIENCE
Another very important notion within Cognitive Linguistics is that of
“entrenchment.” It is defined as “the degree to which the formation and activation of
a cognitive unit is routinised and automated” (Schmid, 2007). However, Schmid
also suggests that entrenchment is, to a certain degree, variable in nature. That is,
despite the viewing of a single lexical item or group thereof as a gestalt – in this case,
an independent unit carrying a specific kind of conceptual information – in the latter
scenario, there may be variation in the use of certain words or their forms depending
on context. As mentioned by Langacker,
When a complex structure coalesces into a unit, its subparts do not thereby cease
to exist or be identifiable as substructures… Its components do become less
salient, however, precisely because the speaker no longer has to attend to them
individually (Langacker, 1987, p.59).
Thus, according to Schmid (2007, p. 121), “it is not only lexical concepts that get
entrenched with repeated use, but also collocational patterns…” The most striking
examples of this kind of structure are idioms, such as piece of cake, break a leg, hit
the books, etc. It is important to note here, however, that the meaning conveyed by
these gestalts is not necessarily equivalent to the sum of the meaning conveyed by its
parts. That is, if viewed from outside the entrenched perspective, such as from that of
a non-native learner of the language (in this case English), the meaning conveyed by
the lexical sub-parts, if put together in exactly the same manner and used within the
same context, may not make sense at all, or would convey something entirely
different. If we take the first example stated here – piece of cake, which means
“extremely easy,” if said in the context
i. That test was a piece of cake,
40
would not make sense to anyone who is not aware of the idiomatic meaning of the
unit. Obviously, a test cannot actually be a piece of cake (in this case an actual
portion taken from a cake), unless of course it is some kind of cake-eating contest, in
which case it may be expressed differently. Although, if the speaker is trying to
create humour through the use of a pun, that is, using the idiom to state that the test
in reality was (to eat) a piece of cake, then the effect would be totally different.
Again, this would be effective only if the hearer is actually aware of the existence of
the idiom and its gestalt meaning. If not, they would (quite figuratively, of course) be
out of their depth.
A bigger study would involve a cross-linguistic approach; however, we gloss over
some examples from Hindi, Gujarati and Marathi so as to elucidate the point further.
As seen below, a similar situation would occur with any of the following idioms in
any of the following languages (contrived examples drawn from native knowledge):
-
Hindi: [ʈu:ʈ pəɖna] (literally - to break and fall) means to attack with all one’s
strength.
i. Example: [nəgaɖe ki aʋaz sunte hi sɪpahi duʃmən pər tu:t pəɖe], means
‘As soon as they heard the drums, the soldiers attacked the enemy
with all their strength.’
-
Gujarati: [ʤəg ʤitʋu] (literally - to win the world), means to achieve absolute
success.
ii. Example: [pərikʃama səfəɭta prapt kərine ene lagju ke eɳe ʤəg ʤiti
lidɦu hətu], means
‘After passing his exams successfully, he felt like he had achieved
absolute success.’
-
Marathi: [ajtjya bɪɭaʋər nagoba] (literally - to be a cobra on a ready-made
burrow), means to declare ownership over someone else’s work; to take credit for
someone else’s work. It also carries the notion of laziness and cunningness, and
41
is possibly drawn from the fact that snakes usually take over holes/burrows
created by other animals.
iii. Example: [səgɦɭə kam toʦ kərto, ti tər tɪthe “saghla kaam toch karto; ti
tar tithe ajtjya bɪɭaʋər nagobaʦ ahe], means
‘He does all the work; she simply sits around and takes the credit.’
Idioms such as these are thus entrenched in a native speaker’s mind as singular,
holistic conceptual units, and while the closed-class elements within them change
according to the context, the rich, conceptual meaning conveyed remains similar
throughout.
Consequently, it can be assumed that non-idiomatic lexical units occurring
together more frequently would also be more entrenched as compared to those that
are less frequent. For instance, the use of stunningly beautiful (frequency score 23)
would be more entrenched than startlingly beautiful (frequency score 6), based on
the corpus data drawn from the British National Corpus through SketchEngine. Thus,
it can be deduced that frequency in usage is directly correlated to the level of
entrenchment.
According to Schmid (2007, p. 121), the “emergence of novel linguistic
structures,” also known as “sanctioning,” can occur due to the facilitation of similar
structures that are already well entrenched. However, he also notes that sanctioning
could be inhibited or blocked by these very structures in certain cases, such as if the
semantic value of both structures is almost congruent. He gives an instance of this as
the use of *stealer instead of thief in English. As the latter is more entrenched, it
seems to block the use of the former as an equivalent substitute.
While this may be true to a certain extent, it must also be noted that sometimes,
especially in cases of the Tip-of-Tongue phenomenon, new structures such as
*stealer can be created and used to substitute the target word, and once established,
could be used more frequently until it becomes a well-entrenched unit in itself. This
process may not be very common, but it cannot be discounted either. A usage-based
example for this very instance is seen in the American sitcom Friends {season 7,
42
episode 11} (2001), wherein Rachel finds out that Chandler is eating a cheesecake
that was meant for their neighbour and calls him a thief, but when she joins him, she
says that they are *dessert-stealers, not dessert-thieves. This latter expression could
have been used for a number of reasons; even the notion that *stealer in this scenario
sounds more aesthetically interesting than thief, although that is entirely a matter of
perspective. Either way, given the popularity of the show and the fact that most fans
still watch it every now and then despite it being over ten years since it stopped
airing, it is highly probable that this term has been entrenched into their minds.
Consequently, it is possible that the word *stealer be used alongside other nouns,
until eventually it becomes an autonomous equivalent for thief. This point is
discussed further in the next section.
We now come to the concept of “salience.” According to Schmid (2007, p. 122),
the theory of spreading activation suggests that “many more words than those that
are uttered in a given speech act are activated during the process of lexical
retrieval.” This implies that in terms of the LCCM Theory, a number of Lexical
Concepts that are associated with the Cognitive Models related to the concept to be
expressed, are activated to a certain degree, and the ones that seem most specific and
apt for the context in linguistic and extra-linguistic terms would form the “active
node” of that conceptual structure. This includes potential “synonyms,” antonyms,
various levels of categorization (superordinate, subordinate etc.) (Rosch, 1999), as
well as collocates and elements within one (conceptual) frame (Schmid, 2007). This
point creates a further basis for the arguments given in chapter 4, and hence shall be
continued there.
3.8 ENTRENCHMENT AND EQUIVALENCE – A DEBATABLE DICHOTOMY
Further to the point of sanctioning as given above, there is a certain degree of
conflict between the concepts of entrenchment and equivalence, posing the question
– What would be the status of equivalence substitution in case of entrenched
collocations and their respective conceptual structures? The answer to this is the
notion of flexibility; human language use and the conceptual system associated with
it are not in any way rigid – they constantly evolve. Given that, we suggest that while
43
entrenchment (especially with respect to collocations) would certainly take the upper
hand in most usage-based situations, in others that demand creativity - which is one
of the distinctive features of human language (Gomez, 1998), such as TOT or even
the writing of fiction or poetry, equivalence substitution can also occur quite
frequently. An instance of this can be seen in lines 15 and 16 of William
Wordsworth’s renowned poem “I Wandered Lonely as a Cloud” (The Wordsworth
Poetry Library, 1994), which is as follows:
i) “The waves beside them danced; but they
Out-did the sparkling waves in glee:
A poet could not but be gay,
In such a jocund company:”
Here, the use of the words gay and jocund could offer similar semantic values on
account of being “synonymous” (Princeton University, 2016) and if not for the
aspects of metre and other such tools employed in poetry, they could easily be
interchanged and still provide the same broad semantic value in both contexts. If
used in the context of company, both gay company and jocund company would create
an almost congruent effect, in terms of pure conceptual detail, as they would in the
“poet…gay” versus “poet…jocund” scenario. The fact that the latter example of
interchanging does not seem appropriate is secondary here. The fact that the general
semantic value is conveyed is of the essence.
Thus, within the Theoretical Background, we have discussed the models of
Language Production suggested by Levelt & Schriefers (1987) and Dell, et al.
(1999), Evans’ LCCM Theory (2009), and the concepts of synonymy, cognitive
economy, equivalence, entrenchment and salience, as well as the relationship
between entrenchment and intralingual equivalence in detail. We have also
established that despite there being seemingly valid points on the functioning of
Lexical Access as a process in the theories suggested by Levelt & Schriefers and
Dell, et al., they still view the phenomenon through the modular perspective, while
also overlooking the factors affecting it. The LCCM Theory, on the other hand,
44
focuses on comprehension rather than production, which leads us to try and address
the objective of bringing it together in a cohesive manner with the other aspects that
have been discussed above. We do this in the following chapter, creating the
foundation for a more streamlined model of Lexical Access in Language Production,
and suggesting the addition of certain points and processes within the same.
4 CMLC – A FUNCTIONAL LEXICAL ACCESS MODEL?
This chapter deals with an amalgamation of all the aspects mentioned in the previous
one, and proposing a functional model of Lexical Access that focuses more on the
factors affecting the processes, and not just the existence thereof. It shall take into
consideration examples from English and compare them with those from other
languages, in order to substantiate the claims made.
4.1 LCCM UPSIDE-DOWN: CMLC?
Given that one of the main objectives of this study is to examine the LCCM
Theory as a functional theory/model of Lexical Access in Language Production, we
shall begin this chapter with the establishment of the LCCM model placed upsidedown – that is, rather than from the ‘Access Semantics through Lexical Input’ point
of view, we focus on the ‘Lexical Access through Semantic Input’ perspective.
We start with the Semantic Potential here – namely, the Cognitive Models within a
speaker’s conceptual repertoire. The argument that Evans posits here, which
indicates the existence of cognitive information created through embodied
experience, seems likely even from the perspective of production. To recap, when it
comes to comprehension, the Cognitive Models associated with a Lexical Concept
come together to form a Cognitive Model Profile for the same, facilitated by the
Access Site associated with the Lexical Concept, based on the linguistic and extralinguistic context of the utterance. We shall now see if the same occurs from a
production point of view, by describing a backwards overview the process, and
seeing if the Cognitive Model Profile provides access to a specific Lexical Concept
45
and consequently, a specific vehicle as well. The system would thus (hypothetically)
work as follows.
i. A set of Cognitive Models would be selected from the overall conceptual
system in order to create associations relevant to the concept to be
expressed.
ii. A Cognitive Model Profile would then be created with reference to the
concept, along with the contextually related and unrelated information
associated with the concept.
iii. The Cognitive Model structure in phase 1 would lead to the Access Site
for a particular Lexical Concept.
iv. This Access Site would facilitate access to the respective Lexical Concept,
which would in turn provide access to the vehicle(s) associated with it.
v. The specific vehicle(s) accessed shall then be used within the utterance.
There seem to be a number of issues in this process. For one thing, while the
narrowing down of the relevant Cognitive Model Profile into a specific set of
conceptual structures based on context seems likely from the comprehension point of
view, it seems rather unnecessary when it comes to production. After all, the speaker
is usually very specific of the exact kind of information he or she wants to express,
and thus, the involvement of a Cognitive Model Profile at this stage (phase 2) seems
redundant. For instance, if a speaker wishes to express the conceptual information
brought together for the notion of shoe, it would be redundant for them to create a
complete, comprehensive conceptual structure with regards to the kinds of footwear
they have seen or worn, and then narrow down from that to create a Cognitive Model
Profile, especially if it involves perceptual input. This would naturally depend greatly
on Context.
Secondly, if each Lexical Concept has a specific Cognitive Model Profile or
Semantic Potential associated with it, then the absence of the latter would suggest the
availability of a number of Lexical Concepts associated with the Cognitive Models
involved, each with their own Semantic Potential which overlap to a great degree.
This consequently suggests that there would be a number of vehicles associated
with the array of Lexical Concepts, which would then require a set of formal and
46
semantic “selectional tendencies” as Evans likes to put it, or a set of factors affecting
the selection of a specific one, in order to carry out the actual selection, while doing
away with the peripheral alternatives.
Thus, while the presence of a Cognitive Model Profile works out well enough
from the comprehension point of view, the very fact that we as speakers are
constantly looking for more contextually apt ways to express ourselves suggests the
existence of more than one Lexical Concept at a time. If the speaker starts with a
specific Cognitive Model Profile which is associated with just one Lexical Concept,
as is indicated, then it is not possible for that to occur, unless a single Lexical
Concept can offer a number of vehicles from which to choose. However, that is not
the case from Evans’ perspective, and hence, the theory needs a certain degree of
modification when it comes to the activation of a Cognitive Model Profile that is
specific to just one Lexical Concept, which in turn provides access to a single
vehicle.
Let us elucidate with the following contrived example.
i. He saw a _____ meadow, in the middle of which stood a ______
structure made of _____ bricks and stones.
Consider the utterance above. Let us focus only on the expression of the
adjectives – say those related to largeness, which the speaker would have to fill up
within the utterance. While he or she could always use the same word to describe all
three nouns, we could place a constraint on doing so, expecting them to use different
words in each blank. Depending on the given context, the Cognitive Models related
to what the speaker intends to express, vis-à-vis the greater-than-average size or
extent of the entities to be described, would be selected from the conceptual system.
Based on these Cognitive Models, the Lexical Concepts associated with each of them
would come in play, and the most appropriate ones would be selected. These would
then give rise to potential vehicles, and the most apt ones would be used within the
utterance. For instance, if the Cognitive Model for the meadow is to be considered,
then the concepts of green, open, grassy, beautiful, vast, and others such as these
would come to mind. Given that the speaker wishes to define the large area covered
by the meadow, he or she would possibly select the Cognitive Models associated
47
with vast(ness), which would then lead to the Lexical Concept [VAST], and hence,
the vehicle vast. This would work in a similar manner in the case of colossal for the
structure and enormous for the bricks and stones.
All the adjectives given above are related to the concept of size, and more
specifically, to a size much greater than would be considered ‘normal’ or ‘average.’
We take the definitions given by Wordnet (Princeton University, 2016) for each of
these words, along with their “synonyms.”
(b)
Vast: (huge, immense) – Unusually great in size or amount or degree or
especially extent or scope. Top 3 nominal collocations from SketchEngine (BNC):
majority, amount and number.
(c)
Colossal: (prodigious, stupendous) – So great in size or force or extent as
to elicit awe. Top 3 nominal collocations from SketchEngine (BNC): statue, cheek
and close-out.
(d)
Enormous: (tremendous) – Extraordinarily large in size or extent or
amount or power or degree. Top 3 nominal collocations from SketchEngine (BNC):
amount, number and pressure.
It seems that vast and enormous have a certain degree of commonness not only in
the semantic value, but also in terms of collocations, as in amount and number.
However, while these collocations may not be the most obvious nominal choices for
colossal, according to the BNC, they do still occur, albeit rarely – twice for
number(s) and 11 times for amount. This ...
Purchase answer to see full
attachment