2. MENTAL LANGUAGE AS AN IDEAL LANGUAGE
3
present when a term immediately signifies several distinct concepts.
These semantic features of Mental Language, which it has in virtue of lit-
erally being the language of thought, make it a powerful language. Modern
interpreters have further attributed a second thesis to Buridan, namely that
the language of thought, Mental Language, is a logically ideal or canonical
language—roughly, that the language of thought is first-order logic, as we
might put it.
3
This daring hypothesis seems attractive on any number of
counts. First, Mental Language as described by Buridan has nearly all the
features required by an ideal language. Second, it identifies the key elements
of a (descriptive) cognitive psychology as concepts and mental operations
on concepts. Third, it provides a framework in which to pursue both logic
and psychology as natural sciences.
Attractive as it may be, I now do not think that Buridan holds that
Mental Language is logically ideal. To see why, let me first develop the
thesis in some detail (§§2–4), and then trace its downfall (§5) on both the
semantic and psychological fronts.
2. Mental Language as an Ideal Language
Mental Language seems to have two further features in addition to (i )–
(iii ), which, when combined with them, render it a logically ideal or canon-
ical language: (iv ) non-redundancy; (v ) logical perspicuousness. Each de-
serves a closer look.
The non-redundancy of Mental Language is a matter of its not contain-
ing any synonyms—or, to put the point a different way, inscriptions and
utterances are synonymous if and only if they immediately signify the same
nibus passiones animae, sicut etiam res quarum ipsae sunt similitudines, ut habetur
primo Peri Hermeneias. Unde ego dico quod propositio mentalis correspondens huic
propositioni, prout est uera, Homo est species non est propositio in qua subiicitur
conceptus specificus hominis, sed est propositio in qua subiicitur conceptus quo con-
cipitur specificus hominis, et ille non supponit pro se, sed pro conceptu specifico
hominis. Ex quo satis patet quod praedicti paralogismi secundum talem mutationem
suppositionum pertinent ad fallacias in dictione.” Buridan justifies this rule by his
realism about Mental as the language of thought, where having a concept in mind just
is to be thinking about what the concept is the natural likeness of. For Jones to have
a concept that doesn’t stand for that of which it is the natural likeness would be for
him not to think about what he is thinking about, which is impossible.
3
This hypothesis was first formulated in Trentman [1970] with regard to William of
Ockham. Trentman does not consider all the characteristics of Mental listed here, but
they seem necessary for Mental to be an ideal language; see the introduction to King
[1985].
c Peter King, unpublished
4
BETWEEN LOGIC AND PSYCHOLOGY
concept(s).
4
Let ‘vixen’ and ‘female fox’ be exact synonyms. As inscriptions
they are distinguished by their orthography, as utterances by their phonet-
ics. As concepts, however, there is no discriminatory medium by which to
tell them apart, so long as the concept itself encapsulates the meaning of
each expression—unless perhaps logical structure does the job: ‘female fox’
seems to be logically complex, composed of parts that immediately signify,
respectively, the concept female and the concept fox. Buridan accepts the
intuition behind this proposal, namely that Mental Language includes ex-
pressions literally composed out of simpler expressions. (This is implicit
in the conception of thought as a language with a vocabulary, syntax, and
formation-rules.) Yet he denies that internal logical structure of this sort
has the same signification as the expression to which it is supposed to be
equivalent, so that ‘female fox’ and ‘vixen’ are not synonymous. Instead,
Buridan endorses the ‘Additive Principle’ for the signification of logically
complex expressions:
5
The signification of a complex expression is the sum of the signifi-
cation of its non-logical terms.
The signification of ‘female fox,’ if the immediate signification is piecemeal,
is the sum of the concept female and the concept fox —all females, foxes
or not, and all foxes, female or not. Thus if Mental contains a concept
corresponding to ‘vixen,’ it need not have any other expression that has the
same signification, and the logically complex expression ‘female fox’ doesn’t
have the same signification as ‘vixen.’ (It stands for vixens, but that is
a matter of supposition rather than signification.) Thus Mental Language
does not include co-significative expressions. Just as it has no room for
ambiguity, it has no room for redundancy.
The last feature that would make Mental an ideal language is logical
perspicuousness. Consider two expressions with the same non-logical con-
stituents, such as “Socrates is taller than Plato” and “Plato is taller than
Socrates.”
6
Unlike written and spoken languages, Mental Language has no
discriminatory medium. Hence the difference between these two expressions
must be explained as a product of the behavior of their logical constituents.
Here ‘is taller than’ is sensitive to order. Hence logical operations in Men-
4
Immediate signification plays the role in Buridan’s system that translation-rules play
in ours, namely correlating expressions in some ‘ordinary’ language with their per-
spicuous canonical representations.
5
Buridan states the Additive Principle in e. g. SDD 4.2.3, Soph. 2 Thesis 5, QM 5.14
fol. 23vb.
6
Nothing rides on these being sentences; the same point could be made with “Socrates
and Plato” and “Socrates or Plato.”
c Peter King, unpublished