It is impossible to trace here the history of those issues, but it is probable that the age-old
vicinity of grammar and logic within the medieval trivium
made much for their unity. The two fundamental disciplines followed one another among the first
school courses, rhetoric being studied last, while hermeneutics was set aside for doctors.
Let us retain however that the two problems are particularly relevant for cognitive research,
because language and thought have always been conceived in a related if not identical way. That
seems to be constant, and enduring theories of an inner language attest to it: whether Platonic
dianoia or logos endiathétos of the Stoics, Augustine's verbum
mentis, William of Occam's lingua mentalis, up until Husserl's
endophasia and Fodor's mentalese.1
The parallel between language and thought has not lost any topicality, if we are to believe
Récanati: "For the last ten years or so, we have witnessed a massive turning of philosophy of
language toward philosophy of the mind [...]. The idea that thought looks very much like a language
has gained pride of place as a central topic in philosophy of the mind" (1991, p.137).
Whereas the rationality of thought (reason being the very form of the soul) has been
admired since saint Thomas, the irregularity of natural languages has unwaveringly been deemed
deplorable, so much so that some even doubt that they can adequately represent
knowledge.2 The unity between language and
could be threatened by the discrepancy. As a consequence, one might say that, among our
contemporaries, traditional logocentrism has grown into logicocentrism, insofar as the inner language
has been suddenly regarded as a formal one. Thus the relationship between the language of thought
and natural languages goes beyond sciences of the language and becomes a concern for the whole of
As far as the relationship between language and thought is concerned, the two problems
require complementary strategies.
(i) The logico-grammatical questioning sets its agenda to folding the linguistic dimension
back onto logic, either in a narrow way, through translations of linguistic fragments into formal
language (as illustrated by Montague and Kamp, in particular, but already criticized by Quine); or
in an "loose" way, with theories such as that of natural logic (Grize), argumentation in language
(Anscombre and Ducrot), relevance (Sperber and Wilson). At the level of signs, the tendency is then
to assimilate linguistic signs and logical symbols (for example, the pragmatics of connectors regards
some grammemes as logical operators).
(ii) The rhetorical-hermeneutic questioning formulates no rationalist assumption on the nature
of thought, and besides there is no consensus at all about the existence of a language of thought. At
any rate, it results in distinguishing between logical and linguistic levels, while insisting on the
inherent heterogeneity of the logic and linguistic dimensions, for logical languages were precisely
instituted to redeem the alleged defects of languages and are put in any case to a great many other
uses. This state of affairs resulted in drawing a fine line between the linguistic sign and the logical
symbol, and correlatively between the hermeneutic modes those signs embody.
Note: By underlining some limitations of logicism, we do not intend to throw
away logic with the logicist bath water. One can consider it regrettable that in this field cognitive
research has by and large kept a low profile, being mainly satisfied with first-order predicative logic
supplemented by a smattering of modal operators, while at the same time research in logic was
deploying an unprecedented variety of formalisms.
1. The thesis of a language specializing in thought and
thought only results from a reworking of Aristotle's Peri hermeneias by late
neo-platonists (by Boethius in particular). [RETURN]
2. The problems of representing knowledge primarily consist
in providing a formalization or at least a logical formulation to segments of scientific or technical
texts: it is in particular a question of dissipating ambiguities arising from linguistic expressions. [RETURN]