Quine and Sellars on Meaning
Ozge Ekin
January 15, 2004
In this paper I intend to examine Sellars’s notion of meaning as it is revealed in his paper “Meaning as Functional Analysis”1. My aim is first to give definitions of terms he uses since Sellars introduces many terms to develop his thesis. Then I will present his analysis on meaning and abstract terms. Finally I will analyze Sellars’ and Quine’s approaches to the matter comparatively.
Let me first establish Sellars intention in
this paper. His aim in “Meaning as Functional Analysis” is to put forward
verbal behaviorism as a tool for understanding aspects of thought and
language. He uses verbal behaviorism as
a tool for dealing with basic concepts of semantical theory. Moreover he
believes that verbal behaviorism can only provide a good theory for coarse-grained
explanatory framework. Hence Sellars’s rules for theory of language
presuppose “rules pertaining to well formedness and relation of deep structure
to surface structure” (p. 457).
Sellars, while giving his classification on
meaning, presents Harman’s analysis in “Three Levels of Meaning”2 which
distinguishes three approaches on “what it is for linguistic expression to have
meaning.” First approach, which Harman
calls a theory of level 1, presupposes that
language is the medium in which we think. Second one which is level 2, is communication. Finally level 3 is on stating and
promising. According to Harman each theory presupposes the former one and a
theory of one level does not give a good theory for the other. In these
theories, Sellars believes that, there is a reference to ‘inner conceptual
episodes’ which are features for fine-grained
psychological explanations and these inner episodes are ‘verbal’ in
analogical sense. Here fine-grained framework is introduced as being developed
on coarse-grained framework where
coarse-grained framework is defined as post-Rylean language. Users of this
language are not aware of beliefs, desires and psychological attitudes hence
they cannot talk about these mental states and they do not have the concept of
thought. Then fine-grained framework starts with Jones in Sellars famous theory
and Jones stipulates that there are inner episodes; accordingly talks about
these inner episodes can start since they initiate certain processes and
awareness of inner episodes brings out
verbal utterances.
Sellars builds his theory on “level 1 theory
of meaning” and refer to what Harman calls ‘thinking in words’ as
thinking-out-loud. He introduces Verbal
Behaviorism as: “On the assumption that such a proto-psychological
framework can be isolated, I shall present in the guise of a claim that
thinking at characteristically human level simply is what is described by this framework. I shall refer this claim as
Verbal Behaviorism.”(Sellars, p.418) He defines, according to VB, thinking
‘that-p’ as “having the thought occur to one that-p has as its primary sense saying ‘p’, and a secondary sense in which it stands for a
short term proximate propensity to say ‘p’.”(Sellars, p.419) Sellars treats
that-clauses as quoted expressions. In such wise
the
thought that 7>5 occurred to Seda
becomes
Seda said ‘7>5’
However, quoting with indirect discourse is
only limited by only speaker’s own language and even with respect to the same
language, different users of the language can make the utterances ‘p’, ‘q’, ‘r’
and these utterances can be described as saying that-p. For clarification of
this matter Sellars suggests that we must exhibit similarity of meaning and its
relation to indirect discourse.
His analysis of meaning starts at that point
and he reveals his hypothesis as: “To say
what a person says, or more generally,
to say what kind of utterance
says, is to give a functional classification of the utterance” .(Sellars,
p.421) Sellars suggests that one’s utterances can have their functioning in the
language according to assent and dissent of the trainers. In this case a child
can learn how one linguistic expression function by her parents help and her
parents know the rules which govern the correct
functioning of the language. He also
points out that “The language learner begins by conforming to these rules without grasping them himself.” (Sellars,
p423) Building on this analysis he
introduces meaning as functional classification in terms of semantic expression
marking contexts within which structurally distinct utterings or inscribings
what he calls “natural-linguistic objects” are classified in terms of their
functions in language entry transitions,
language departure transition and intra-linguistic
moves. He defines these three pattern governed linguistic behavior as
follows:
Language entry transitions are linguistic
responses to perceptual stimuli, language departure transitions are the speaker’s
responds to linguistic conceptual episodes as non-linguistic conducts. Lastly intra-linguistic
moves are defined as the speakers linguistic conceptual episodes tend to occur
in patterns of valid inference, and tend not to occur in patterns which violate
logical principles i.e. inferential transitions from one linguistic expression
representing to another. Hence ‘means’ does not only give a description or
equate simply one linguistic term with another. It is rather a specialized form
of the copula. Therefore Sellars, correctly, argues that we cannot use ordinary
quotation since it lacks the match of this special copula, consequently he
presents illustrative sortals to
represent functional modes of sorting. In the light of this argument we can
start to analyze meaning as a copula:
Consider
‘Und’ (in German) means and.
According to Sellars since subject of this
sentence is a singular term and there is an unusual use of the word ‘and’ there
are many ambiguities arise analyzing this expression such as taking ‘und’ as a
universal. However there are many ‘und’s in German functions in different ways.
Moreover this usage does not tell us that ‘und’ and ‘and’ have the same
meaning, it gives the meaning. Hence for functional purposes by naming singular
terms as distributive singular terms he
suggests that we should use:
‘Und’s
(in German) are ·and·s
where to be an ·and· is to be an item in German
which functions as ‘and’ does in English.
I believe, Sellars, introducing this
classification resolves an important ambiguity in use and mention conflict. While doing this he also relates the
classification of the truth context. In his usage, abstract terms become
distributive singular terms true of
appropriately concatenated property such as:
Triangularity is true of a
tells us
Expressions consisting of ·triangular· appropriately concatenated
with and ·a· are true.
In accordance, he correctly suggests that,
nominalizing devices such as ‘-ity’, ‘-hood’,
‘-ness’ form corresponding abstract singular
terms and define metalinguistic functional sortals. Moreover they turn them
into distributive singular terms. However, Sellars, while giving the definition
of “having the same meaning” in
translation for an abstract term, oversimplifies the matter. In this paper (Meaning as Functional Classification)
his examples are only about connectives and mathematical objects. Furthermore, even his suggestion is valid for
these terms it does not work for long
sentences as Kripke points it out; and this, Sellars argues, is because,
distinction of attributes and classes is not clearly understood . Here Quine
suggests that even though we do not draw
this distinction we can make long expression in dot quotes work as a predicate
of another expression by simply analyzing expression as sequence of the
successive simple components.
Quine and Sellars share the similar ideas on abstract terms. Sellars argues that abstract singular terms are not labels of irreducible eternal objects but metalinguistic distributive singular terms where Quine believes that there is no object one can name for example as “redness” but this singular term has a meaning. Sellars defines having the same meaning as having the same battery of rules or rather having the same functional roles in a language L without talking of an extra-linguistic world. He defines meaningfulness purely by means of inferential mechanism. However Quine argues that there cannot be any distinction between matter of meaning and matters of fact. He unlike Sellars disputes that truth of any statement depends upon the meanings of the words used and on extra-linguistic fact. Quine, consequently, concludes that meaning cannot make an isolable contribution to the truth of the statement. I do not agree with Quine in this view. I believe Sellars’ account of meaning provides a sound theory as without extra-linguistic world we can decide the truth of a statement. As in Burge’s example3, having an the incomplete understanding of arthritis, a patient thinks that he has an arthritis in his thigh, but this may not be the case of incomplete understanding in another social environment given that no physical fact changed. Hence it is only our inferential mechanism that is learned in a functional role changes the meaning of a word.