=Paper=
{{Paper
|id=Vol-1419/paper0089
|storemode=property
|title=The Relation between Cognitive and Linguistic Structures
|pdfUrl=https://ceur-ws.org/Vol-1419/paper0089.pdf
|volume=Vol-1419
|dblpUrl=https://dblp.org/rec/conf/eapcogsci/VosgerauP15
}}
==The Relation between Cognitive and Linguistic Structures==
https://ceur-ws.org/Vol-1419/paper0089.pdf
The relation between cognitive and linguistic structures
Gottfried Vosgerau (vosgerau@phil.hhu.de)
Department of Philosophy, Heinrich-Heine-Universität Düsseldorf
Universitätsstr. 1, 40225 Düsseldorf, Germany
Wiebke Petersen (petersen@phil.hhu.de)
Department of Computer Linguistics, Heinrich-Heine-Universität Düsseldorf
Universitätsstr. 1, 40225 Düsseldorf, Germany
Abstract 1957). It postulates five levels which need to be ultimately
accounted for by a comprehensive theory of linguistic mean-
While there is large agreement that there must be a systematic
relation between the cognition of single individuals and the ing:
public meaning of lingusitic expressions, there is no theory that
could describe this relation systematically and formally. The (5) recursive semantics for the object language L
aim of this paper is to describe a framework that is able to fill
this gap. We will use frames as an adequate format to describe
both mental representations and linguistic meaning. The cen- (4) theory of intersubjective, conventional meaning of utter-
tral thesis is that the rich structures of individuals’ representa- ances (what is said)
tions overlap in the sense that they share a common core. This
core can be identified as the public meaning of the word asso- (3) theory of subjective meaning of utterances (what is meant
ciated with these mental representations (concepts). Both lev-
els are systematically related by abstraction mechanisms (from / what is understood)
cognition to language) and attunement mechanisms (from lan-
guage to cognition). (2) psychological theory of the mental states (propositional at-
Keywords: concepts; frames; meaning; mental representa- titudes) of the users of L
tion; natural language; communication
(1) description of the observable behavior of the users of L.
Linguistic Meaning and Individuals’ Cognition
Although serious doubts have been raised against the de-
One of the most central questions in the philosophy of lan-
tails of the program (e.g. against the possibility of describing
guage is the metaphysical question what meaning is and how
propositional attitudes independent of conventional meaning;
it is constituted (or, to use the title of a seminal paper by Put-
see Burge 1979), the overall rationale of the framework is
nam, what the “meaning of ‘meaning”’ is; Putnam 1975).
still widely accepted (for a cognitive expansion of the general
The most influential (and traditional) view is that meanings
framework see Sperber and Wilson 1986). However, one of
are some kind of objective, abstract entities or relations which
the general problems is relating the intersubjective meaning
are shared or commonly accessed by speakers of a language
(henceforth the public meaning) on level (4) to the content
community (see, e.g., the classical arguments in Frege 1892).
of individual mental representations on level (2). Although
However, language users need some “grasp” or representa-
level (3) might in principle seem to provide a link, the usual
tion of such meanings in order to communicate. Although it
description of this level in the Gricean tradition is, first, con-
is widely agreed that the relation between the individual level
cerned with a rather inter-subjective level of communication,
of representing entities (objects, situations, etc.) and the pub-
which is not thoroughly linked to a truly subjective level of
lic level of linguistic meaning has to be described in detail, it
representing entities, and is, second, restricted to the mean-
has mainly been described as a problem: For example, in the
ing of utterances of whole sentences. This paper, however,
Fregean tradition the question is what it is to grasp a meaning
is mainly focusing on the meaning of smaller units such as
(a “thought” or a “sense” in the Fregean terminology), i.e.
phrases and words on the linguistic level and concepts on the
how the abstract meanings are reflected in individual repre-
mental level, for which the link between public meaning and
sentations of entities. In the (later) Wittgensteinian tradition
individual mental representations has rarely been systemati-
of use-based theories, the question arises how it is possible to
cally discussed. Thus, although the paper follows the general
learn the public meaning on the basis of individually different
idea of the Gricean program, it will not assume that the rela-
sets of uses one is acquainted with. In recent “cognitive se-
tion between public meaning and individual mental represen-
mantic” theories, which take public meanings to be strongly
tations has to be described as a three-step relation in the way
based in individual cognition, the problem is to explain the
Grice describes it.
emergence of linguistic meanings from disparate and maybe
In order to develop such a systematic account of the re-
inhomogeneous individual meaning representations.
lation, we need three things: First, we need a detailed and
The most influential framework to address such questions
is probably the so-called “Gricean Program”1 (cf. Grice, theory of implicatures; rather, the latter is Grice’s own way of car-
rying out the program, which can also be carried out in many non-
1 The “Gricean Program” as such does not contain the Gricean Gricean ways.
542
adequate format of describing individuals’ mental represen- value forest of the attribute LOCATION can be further spec-
tations that is rich enough to account for the individual dif- ified by attributes like SIZE or TREE SPECIES. Due to their
ferences. Second, we need a detailed and adequate format recursivity, frames are flexible enough to represent informa-
for describing the public meaning of linguistic expressions. tion of any desired grade of detail. We assume that attributes
Third, we need a systematic link between the two descrip- in frames assign unique values to objects and thus describe
tions, one which explains how public meanings “emerge” functional relations. Formally, frames can be represented by
from mental representations of multiple individuals, i.e. that connected directed graphs where the arcs correspond to at-
explains how public meaning is “grounded” in mental repre- tributes. As attributes are functions, no node may have two
sentations. equally labeled outgoing arcs. The nodes may be labeled by
This paper aims at sketching a first frame-work for such a types which restrict the attribute domains and ranges, i.e. the
systematic description of the relation between public mean- set of objects for which an attribute is adequate and the set of
ing and individuals’ mental representations. It assumes that values an attribute can take.
frames constitute a suitable format of describing both men-
tal representations and public meaning. If, as assumed here,
e round
both levels can be described with the same format, the rela- sh a p
tion between the two can be described as a relation between
structures in the same format, which renders the whole en- shelter l oc
atio
terprise feasible. We will now introduce frames as a format n
of describing mental representations, especially mental rep- tree species
forest oak
resentations of classes of entities (concepts2 ). We will then
explicate the central thesis that public meaning is the over-
lap between the structures of the mental representations of Figure 1: Frame of the concept ‘round shelter in oak forest’
different language users, before we will discuss mechanisms
of abstraction, which lead from individuals’ mental represen-
tations to public meaning, and mechanisms of attunement, Figure 1 shows two additional notational devices which we
er
mo
which describe the influence of public meaning on individu- use in our graph representations of frames: First, the double
th
th
border at the shelter node marks it as the central node of the
mo
als’ mental representations.
er
frame; it indicates that the graph represents a frame about
Frames as Structures of Concepts shelters. Second, by using a rectangular node for the cen-
Barsalou (1992) proposes that concepts can be naturally de- tral shelter node, we identify the whole frame as a concept
scribed in terms of frames. Frames as recursive attribute- or category frame which corresponds to a 1-place predicate
value structures have been widely used as a general for- the argument of which is represented by the shelter node. A
mat for knowledge representation, e.g. for capturing lin- round central node would indicate instead that it is a frame
guistic knowledge (Fillmore, 1970) or situational knowledge of a not further specified category member of the category
(Minsky, 1975). Advancing the basic ideas of Minsky (1975), ‘round shelter in an oak forest’.
Barsalou (1992, p. 21) argues for frames as “dynamic rela- In contrast to other frame theories, our frames are capable
tional structures whose form is flexible and context depen- of representing not only sortal concepts like SHELTER, which
dent”. He presents psychological evidence for attribute-value denote classical categories, but also relational ones like SIB -
structures derived from behavioral animal studies. These LING or MOTHER (Petersen, 2007); the referents of the lat-
studies indicate that animals encode stimulus information as ter concepts are given by a relation to a possessor (“sibling
attribute values and not as independent features. Furthermore, of”, “mother of”). Frames of relational concepts differ from
he gives empirical evidence for the importance of conceptual frames of sortal concepts in that they have an additional rect-
relations in human cognition. Here, we will briefly sketch our angular node for the possessor argument. Figure 2 shows the
frame account which builds on Barsalou’s flexible cognitive frame for the sibling concept. It consists of three nodes, one
approach, but provides it with a rigid formal foundation (see for the sibling itself (rectangular, double border), one for the
also Vosgerau, Seuchter, & Petersen, in press). person it is the sibling of (rectangular, single border) and one
The attributes in a concept frame are the general properties for the mother of both (round). The relation between the two
or aspects by which the respective concept is described (e.g., persons is constituted by the fact that they both have the same
SHAPE, LOCATION).3 Their values are concrete or under- mother.4 This is modeled by the single node to which the two
specified specifications (e.g., [SHAPE: round], [LOCATION: MOTHER-arcs point. Note that in contrast to classical frame
forest]). The attribute values can themselves be complex accounts our approach does not presuppose that the central
frames and thus described by additional attributes. E.g., the node of a frame, i.e. the node which determines what is de-
2 Throughout the paper, the term “concept” will be used for men- 4 Note that all frames in this paper are severely simplified. The
tal representations and not for linguistic expressions. frame in figure 2 e.g. models the sibling concept as being purely
3 Throughout this paper attributes are typeset in capitals and their determined by the mother relation, leaving aside fathers or socially
values in italics, while concepts are set in small caps. established family relations.
1
543
shelter l oc
atio
n
tree species
forest oak
noted by a frame, is a root node of the frame graph.5 ically describe and explain various mental abilities in a way
that conforms to the experimental research in psychology and
related disciplines (cf. Vosgerau, 2009, 2011).6 In partic-
ular, we have developed a cognitive theory of thoughts for-
er
mo
mulated within a representational framework (Vosgerau &
th
th
mo Synofzik, 2010) and shown that our notion of representa-
er
tion is able to integrate ideas of “grounding” and “embod-
iment” (Seuchter & Vosgerau, 2011; Weber & Vosgerau,
2011, 2012), which are sometimes taken to speak against rep-
resentations. By using oscillatory neural networks as a bio-
logically motivated model, Petersen and Werning (2007) give
Figure 2: Frame for the concept ‘sibling’ evidence for the cognitive adequacy of our frame model and
shows how frames might be implemented in the cortex.
An adequate frame theory needs means of restricting the
In particular, frames are flexible enough to describe con-
set of admissible frames. Therefore, frame nodes may be la-
cepts that do not reduce to definitions or feature lists, which
beled by types. The types are ordered in a type hierarchy
are central to the “classical” theory of concepts (e.g. Katz,
which is enriched by appropriateness conditions which con-
1999). That is because it is possible to introduce constraints
strain the domain and range of attributes. Thus the type sig-
into frames that relate, for example, certain attributes to possi-
nature tells which type of entities can have a certain attribute
ble values of other attributes or that determine the presence of
and of which type the values of each attribute are. So, we can
a certain attribute given a certain value. Additionally, frames
infer from an attribute the type of the nodes its arc connects,
seem to be suited to incorporate statistical information on
unless it is further restricted by other constraints in a partic-
attributes and values, which allows adopting specific advan-
ular frame. For example, the type signature specifies that the
tages of prototype theory, namely their description of vague
domain of the attribute MOTHER is person and that its range
concepts. With statistical information and constraints, the in-
is woman. In contrast to figure 2, figure 1 shows an example
formation, e.g., that the typical shape of a tree depends on
of a frame in which the type labels at the nodes effectively
whether the tree has leaves or needles can be incorporated,
restrict the attribute domains and ranges. E.g., LOCATION is
which is not (straightforwardly) possible neither in the clas-
a very general attribute which applies to all kind of physical
sical theory nor in prototype theory. All of these features
objects (not only to shelters) and which takes all kind of lo-
are pivotal for describing the diverseness of individuals’ con-
cations as values (not only forests). The formal details of our
cepts: If a concept would be described in a definition-like
frame account can be found in Petersen (2007) and Petersen
way, e.g., there would be only one “correct” way of repre-
and Osswald (2014).
senting it. In this case, language users would either possess
Although it is in principal possible to express the infor-
the concept or not, and successful communication between
mational content of a frame in classical logical formulas, it
a concept-possessor and a non-possessor could hardly be ex-
is more natural to assume that concepts are mentally stored
plained. Moreover, the different uses of a word a single indi-
in terms of frames than in terms of formulas. As Barsalou
vidual is acquainted with could be reflected in different statis-
(1992) points out, there is empirical evidence for attribute-
tical information in the two individuals’ frames for one con-
value sets and relations in cognition. In frames, concepts re-
cept.
main units although they may be highly structured. This unity
To give a short example: A child, a car mechanic, and a
is concealed in logical formulas by multiple occurring1 vari-
hunter might have very different concepts of cars. While the
ables. Furthermore, as predicates are constants in predicate
child can be expected to have a quite limited representation of
logic, one is forced in a formula representation to stipulate a
technical details, the car mechanic will probably have much
fixed arity and a fixed order of arguments for each relation ex-
more “technical attributes” (such as VALVES as an attribute
pressed by a predicate. Here, frame theory is more flexible:
of the engine) in her frame. The hunter, in contrast, may rep-
As relations are decomposed into functions with relational
resent attributes related to cross-country mobility in his car-
arguments, one gains a parameterized access to relational ar-
frame. In this way, frames can be used to capture the individ-
guments. Thus frames have a non-linear structure in which
uals’ differences in concept representations. However, since
substructures can be addressed via labeled symbols instead
frames are structures, there might still be a “core” of each in-
of ordered argument positions which is cognitively more ad-
dividuals’ frame that is also present in the others’. In the car-
equate a description of mental representations.
example, this core might contain attributes such as COLOR,
Although some critics have tried to eliminate the notion
WHEELS, ENGINE, FUEL.
of representation (e.g. Brooks 1991,Thelen 1995; for a dis-
cussion, see van Gelder 1998), a liberal notion of represen- 6 The liberal notion defines mental representations as internal
tation has been developed and fruitfully used to philosoph- states that stand for something else without presupposing that they
are symbolic, conceptual, comprehensive, or amodal. Since the cri-
5 A root of a directed graph is a node from which all other nodes tique of representations attack one or more of the additional specifi-
can be reached via paths of directed arcs. cations, the liberal notion is not affected by this critique.
544
Overlap Between Frames as Meaning concepts, it is likely that different mechanisms of attunement
within language communities are at work so that the com-
The central thesis of this paper is that the public meaning of
mon use of language also shapes the concepts (for a review
linguistic expressions can be understood as the “overlap” be-
of literature showing the impact of language use on concept
tween the rich structures of individuals’ concepts. Frames
acquisition, cf. Rakoczy 2010, sec. 4); these mechanisms in-
provide a promising tool to accomplish this task: Due to
clude explicit explanations, extra-linguistic reference fixing
their recursive structure, the same semantic unit can be de-
(e.g. through ostention), and implicit alignment through con-
scribed using very different degrees of granularity, depending
textual disambiguation, also involving the “common ground”.
on how much of the values are further specified by attribute-
We will come back to possible attunement mechanisms after
value pairs. This allows the identification of a core structure,
describing the counteracting mechanisms of abstraction and
which is common to all different fine-grained representations
enrichment that are at work in concept acquisition, language
of different individuals (the “overlap”). The core structure
learning, and language use.
can then be conceived of as the abstraction of the individuals’
representations, which constitutes a distinct level of descrip- Abstraction and Enrichment
tion, the level of public meanings.7 We will now spell out
In general, abstraction mechanisms lead from specific repre-
this central idea in more detail and identify open questions
sentations to more general representations, i.e. to concepts.
for further elaboration.
Such mechanisms mark the relation between instantiated
The overlap between the structure of individual represen- frames (standing for particular objects or situations) and gen-
tations can be conceived of the result of the generalization eral frames (standing for types of objects or situations). En-
operation on frames (Carpenter, 1992). On the one hand, richment mechanisms are counteracting abstraction and lead
the computational complexity can be expected to be compar- from general frames to instantiated frames. Such mechanisms
atively low given that for a specific concept or word the cen- are employed if a particular entity is subsumed under a con-
tral node is given; on the other hand, the operation has to be cept, i.e. in categorization: If, for example, an entity is cat-
adapted to handle statistical information and constraints (e.g. egorized as a car, the general car-frame is enriched by spe-
by computing mean values). However, more pressing is the cific attributes and values (e.g. the value red for the color of
question whether such overlaps can be assumed at all, i.e. the the car, or the attribute SPOILER) to represent the specific
question if public meanings do exist. For if there is no such features of this particular car. Since enrichment mechanisms
overlap, communication cannot be successful. First, it should can be expected to be, in general, just the converse of the ac-
be noted that there is no strict boundary between successful cording abstraction mechanisms, we will now concentrate on
communication and non-successful communication, as is also abstraction.
reflected by the notorious difficulties to distinguish between One simple abstraction mechanism eliminates specific val-
different languages, dialects, sociolects, technical languages, ues from frames. To take the same example: If a child en-
etc. (cf. Ethnologue, 2014). In fact, we propose that language counters different cars, she will be able to abstract from the
communities are constituted by the overlap. Take, e.g., youth different colors of cars by eliminating the value of the color-
“language”: whether a given word (e.g. the German adjec- attribute. In this way, the general representation still contains
tive “porno”) is part of youth language is determined by the the information that each car has some color. A related mech-
degree of overlap between young people’s mental representa- anism will assign statistical information to different possible
tions associated with the word and the non-overlap with the values. Such a mechanism would result in general frames
not-so-young people’s representations. Thus, language com- representing prototype-concepts in the sense of Osherson and
munities are not given independent of overlaps. (Of course, if Smith (1981). A further abstraction mechanism will elimi-
one speaker does not know a word and thus has no mental rep- nate whole attributes that are specific for only some exem-
resentation associated with it, there can be no overlap; this is plars of a category. For example, the attribute SPOILER is
most obvious for completely different languages.) Second, it not adequate for every car but only for some. Thus, if a child
can be assumed that there are different mechanisms that con- sees the first car without a spoiler (after having been exposed
tribute to the establishment of overlaps: for basic perception- to different cars with spoilers), she can abstract from this fea-
based concepts, the mere similarity of the perceptual and ture by eliminating the attribute SPOILER in her general car-
cognitive apparatuses in different individuals assures a con- frame (her concept). Such simple abstractions can be formal-
siderable overlap (basic object categorization seems even to ized by generalizations over all representations of particulars
be independent of the first language of a person; cf. Malt, belonging to one category.
Sloman, and Gennari 2003). For more complex or abstract However, the most interesting and probably the most im-
7 Metaphysically speaking, meanings are thus abstract entities, portant abstraction mechanism is the one that introduces new
namely certain structures. Mental representations can have a the attributes. Let us illustrate the basic idea with the example of
same structure as linguistic expressions; in such a case, the content the mother-concept of a child. A young child will acquire a
of the mental representation will be identical to the public meaning concept of his mother only, i.e. he will not (yet) understand
of the linguistic expression. Of course, mental representations and
linguistic expressions have differen “realizations” or “vehicles” and that other people also have a mother. It is likely that the child
are thus never identical. does not explicitly represent himself as a constituent of the
545
mother frame; on the contrary, his mother-frame will only be Attunement
implicitly relate to himself (cf. Seuchter & Vosgerau, 2011). So far, abstraction mechanisms lead us from individual cog-
This means that initially, the mother-frame will not contain nition to public meaning. However, obviously the use of a
a mother-attribute mapping from the children of the mother common language also has a huge influence on the way we
to the mother. Only later, when the child learns that other represent things, especially on our conceptual systems. Thus,
children have different mothers, the child will add this at- the framework has to be complemented with mechanisms that
tribute to his mother-frame and so gains the power of rep- describe the influence of language use on individuals’ con-
resenting different people as the (different) mothers of differ- cepts. We will call these mechanisms “attunement mecha-
ent children. This abstraction mechanism is likely to play a nisms”. The basic idea is that the overlap between the differ-
major role also in adult learning, especially in scientific learn- ent individuals’ frames is increased by attunenment mecha-
ing: Often enough, scientific progress is made by introducing nisms. The advantage of our framework is that such mecha-
more attributes that lead to more abstract, i.e. more general nisms can be described in detail, such that this description can
theories. For example, the identification of light and X-rays also serve as a basis for empirical work. So far (to our knowl-
with electromagnetic waves was only possible by adding the edge), such mechanisms have not been described in detail (at
attributes of electric energy, frequency and wave-length to the least not in any technical detail), such that most work is still to
according concepts. be done. However, we will now sketch some possible attune-
Abstraction and enrichment mechanisms relate the differ- ment mechanisms to illustrate the general idea and to prove
ent levels of frames with each other. Both on the cogni- the feasibility of our framework.
tive and the linguistic level, we find instantiated and general One class of attunement mechanisms can be described as
frames (see also table 1). All four levels play a crucial role implicit, since they do not involve communication about the
meaning of words. One such mechanism is the use of point-
ing gestures along with words. In this way, another speaker
Table 1: Four levels of frames. learns about the extension of the partner’s concept (e.g. the
instantiated general
concept RED) associated with the word and can attune her
individual mental concepts (mental
own concept. Another implicit mechanism involves the con-
speakers’ representation representation
text and the common ground that helps to clarify what is re-
representation of particulars of types)
ferred to. One concrete example: If we are going outside and
public meaning utterance semantic meaning
you utter “This drizzle is annoying” while opening your um-
meaning (in (context-inde-
brella, I will probably understand that you are talking about
specific context) pendent meaning)
the rain even if I don’t know the word “drizzle”. Both at-
tunement mechanisms mentioned help to fix the referent (the
central node), such that the hearer can use the abstraction
in understanding the relation between individual psychology
mechanisms to attune his or her concepts with the speaker’s
and public meaning and their interaction in communication:
concepts. However, there are also implicit attunement mech-
Individuals have representations of concrete particulars with
anisms which do not work through fixing the referent. For
a rich structure of attributes and values. By abstraction, they
example, if I hear someone say “I don’t like boarhounds, I
also form general representations (concepts) of types of en-
prefer long fur”, I can learn that boarhounds have short fur.
tities. Once they are acquired, they can, in turn, be used to
The other class of attunement mechanisms contains explicit
represent particulars, in which case context-specific attributes
mechanisms that involve a kind of “meta-communication”
and values are again added to the representation (“enrich-
about word-meanings. For example, a speaker might give
ment”). Linguistic expressions (words) are associated with
an explicit definition of a term she uses or paraphrase it.
the concepts and can be used to communicate about specific
But meanings can also be bargained about in a communica-
things in specific contexts.8 The utterance-meanings of such
tion, if the communicators have different ideas (e.g. whether
expressions used in specific contexts can be described as the
this shade of green qualifies as teal). Moreover, people can
overlap between the structures of the individuals’ instantiated
refer to expert-knowledge, as in “Philosophers understand
frames. Abstracting from the specific context will result in
this term as referring to ...”. Obviously, explicit attunement
the semantic meaning of a linguistic expression, which can
mechanisms aim at creating a more consistent use of words
be described as the overlap between the structures of individ-
between individuals, which is described as an increase in
uals’ concepts in a given language community.
concept-overlap according to our framework.
8 While full competent speakers mostly have knowledge about
Conclusion
the public core, i.e. the semantic meaning of a word, and often
enough even have specific knowledge about the specific concept of Frames are rich enough to provide a detailed and adequate
the person they talk to, both kinds of knowledge are not necessary description of individuals’ mental representation, in partic-
(although very helpful) for communication and are likely not to be
present in young children who are nevertheless competent commu- ular the representation of concepts. Due to their recursive
nicators. structure, they are apt to describe concepts at different levels
546
of granularity. Therefore, they are also apt for describing the Petersen, W., & Osswald, T. (2014). Concept composition in
public meaning of linguistic expressions. Thus, a systematic frames: Focusing on genitive constructions. In T. Gamer-
explanation of the relation between cognitive structures and schlag, D. Gerland, R. Osswald, & W. Petersen (Eds.),
linguistic meaning becomes possible. We elaborated on the Frames and concept types. applications in language, cog-
thesis that public meaning can be understood as the overlap nition, and philosophy (Vol. 94, pp. 243–266). Heidelberg,
between the individuals’ frames for concrete particulars and New York, Dordrecht, London: Springer.
for classes of entities. According to our framework, the level Petersen, W., & Werning, M. (2007). Conceptual fingerprints:
of public meaning is grounded in the cognitive level by ab- Lexical decomposition by means of frames. In U. Priss,
straction mechanisms, while it has a huge influence on the S. Polovina, & R. Hill (Eds.), Proceedings of the iccs 2007,
cognitive level through attunement mechanisms. In this pa- lnai 4604 (p. 415-428). Berlin: Springer.
per, we could only describe a general framework which leaves Putnam, H. (1975). The meaning of ‘meaning’. In H. Putnam
much detail work for the future, especially regarding the for- (Ed.), Philosophical papers (pp. 215–271). Cambridge and
mal description of attunement mechanisms. New York: Cambridge University Press.
Rakoczy, H. (2010). From thought to language to thought:
References Towards a dialectical picture of the development of think-
ing and speaking. Grazer Philosophische Studien, 81(1),
Barsalou, L. W. (1992). Frames, concepts, and conceptual
77–103.
fields. In A. Lehrer & E. F. Kittay (Eds.), Frames, fields,
Seuchter, T., & Vosgerau, G. (2011). Grounding abstract con-
and contrasts: New essays in semantic and lexical orga-
cepts: Causal indexicals and affordances. In B. Kokinov,
nization (pp. 21–74). Hillsdale: Lawrence Erlbaum Asso-
A. Karmiloff-Smith, & N. J. Nersessian (Eds.), European
ciates.
perspectives on cognitive science: Proceedings of the eu-
Brooks, R. A. (1991). New approaches to robotics. Science,
ropean conference on cognitive science (eurocogsci 2011).
253(5025), 1227–1232.
Sofia: New Bulgarian University Press.
Burge, T. (1979). Individualism and the mental. Midwest
Sperber, D., & Wilson, D. (1986). Relevance: Communica-
Studies in Philosophy, 4(1), 73–121.
tion and cognition. Cambridge and Mass: Harvard Univer-
Carpenter, B. (1992). The logic of typed feature struc- sity Press.
tures: With applications to uunification grammars, logic Thelen, E. (1995). Motor development. a new synthesis. The
programs, and constraint resolution (Vol. 32). Cambridge American Psychologist, 50(2), 79–95.
and New York: Cambridge University Press. van Gelder, T. (1998). The dynamical hypothesis in cognitive
Ethnologue. (2014). The problem of language iden- science. The Behavioral and Brain Sciences, 21(5), 615–
tification. (http://www.ethnologue.com/about/problem- 628.
language-identification; last access: 2015/3/1) Vosgerau, G. (2009). Mental representation and self-
Fillmore, C. J. (1970). The case for case. In E. Bach & consciousness: From basic self-representation to self-
R. T. Harms (Eds.), Universals in linguistic theory (p. 1- related cognition. Paderborn: Mentis.
88). London, New York, Sydney, Toronto: Holt, Rinehart Vosgerau, G. (2011). Varieties of representation. In
and Winston, Inc. A. Newen, A. Bartels, & E.-M. Jung (Eds.), Knowledge
Frege, G. (1892). Über Sinn und Bedeutung. Zeitschrift für and representation (pp. 185–209). Stanford and Calif and
Philosophie und philosophische Kritik, 100, 25–50. Paderborn and Germany: CSLI Publications and Mentis.
Grice, H. P. (1957). Meaning. The Philosophical Review, Vosgerau, G., Seuchter, T., & Petersen, W. (in press). An-
66(3), 377. alyzing concepts in action-frames. In T. Gamerschlag,
Katz, J. (1999). On the general character of semantic the- D. Gerland, R. Osswald, & W. Petersen (Eds.), Meaning,
ory. In E. Margolis & S. Laurence (Eds.), Concepts: Core frames, and conceptual representation, studies in language
readings (pp. 125–150). Cambridge and Mass: MIT Press. and cognition. Düsseldorf: Düsseldorf University Press.
Malt, B. C., Sloman, S. A., & Gennari, S. (2003). Speaking Vosgerau, G., & Synofzik, M. (2010). A cognitive theory of
vs. thinking about objects and actions. In D. Gentner & thoughts. American Philosophical Quarterly, 47(3), 205–
S. Goldin-Meadow (Eds.), Language in mind (pp. 81–111). 222.
Cambridge and Mass: MIT Press. Weber, A., & Vosgerau, G. (2011). Is cognition grounded in
Minsky, M. (1975). A framework for representing knowl- action? conceptual considerations and empirical evidence.
edge. In P. Winston (Ed.), The psychology of computer In B. Kokinov, A. Karmiloff-Smith, & N. J. Nersessian
vision. New York: McGraw-Hill. (Eds.), European perspectives on cognitive science: Pro-
Osherson, D. N., & Smith, E. E. (1981). On the adequacy of ceedings of the european conference on cognitive science
prototype theory as a theory of concepts. Cognition, 9(1), (eurocogsci 2011). Sofia: New Bulgarian University Press.
35–58. Weber, A., & Vosgerau, G. (2012). Grounding action rep-
Petersen, W. (2007). Representation of concepts as frames. resentations. Review of Philosophy and Psychology, 3(1),
The Baltic International Yearbook of Cognition, Logic and 53–69.
Communication, 2, 151-170.
547