Kenny Anthony. Medieval Philosophy: A New History of Western Philosophy. Volume 2
Подождите немного. Документ загружается.
Whether ‘being’ is analogous or univocal is a murky question not because
of diYculties about analogy but because of the almost universal opacity of
the medieval notion of being. If we are talking about existence, as expressed,
say, in the sentence ‘There is a God’, then the question whether being is an
analogous or univocal predicate does not arise since attributing existence to
something is not a matter of attaching a predicate to a subject. But, in Scotus
at least, ‘to be’, period, seems equivalent to a vast disjunction of predicates :
‘to be a horse, or a colour, or a day, or ...’ and so on ad inWnitum. So
understood, ‘to be’ seems clearly univocal. Suppose that there were only
three items in the universe, A, B, and C. The predicate ‘...iseither A, or B,
or C’ seems to attach in exactly the same sense to each of the three items.
Modistic Logic
Scotus did not make any substantial contribution to formal logic, though
his metaphysical ideas on the nature of power and potentiality were to
have a signiWcant long-term eVect on modal logic. He was, however, long
credited with an interesting work on the borders of logic and linguistics, a
Grammatica Speculativa that the young Martin Heidegger took as the subject of
his doctoral thesi s. The work is now regarded as inauthentic by scholars,
and attributed not to Scotus but to his little-known contemporary Thomas
of Erfurt, writing about 1300.
The work is important as representative of a new approach to logic,
adopted by Radulphus Brito (d. 1320) and a number of thinkers in the late
thirteenth century, known as ‘modistic logic’ in contrast to the ‘terminist
logic’ which we have seen in the works of Peter of Spain and William
Sherwood. Rather than studying the properties of individual terms, these
modist logicians studied general grammatical categories—nouns, verbs,
cases, and tenses, for insta nce—which they called modi signiWcandi, or ways
of signifying.
Meaning, according to the modists, was conferred on sounds by human
convention, which they called ‘imposition’. The unit element of meaning
was the dictio, ‘diction’. A single diction might embrace many diVerent
verbal forms: the cases of a Latin noun, for instance, plus the adjectives and
adverbs associated with it. A favourite example was the diction for pain,
which included the noun ‘dolor’ in its diVerent cases, the verb for feeling
LOGIC AND LANGUAGE
142
pain ‘doleo’, and the adverb ‘dolenter’, meaning ‘painfully’. The basic
convention setting up the diction for pain was called by the modists Wrst
imposition; further conventions, by a second imposition, established these
modi signiWcandi that linked diVerent word forms to diVerent types of use.13
Some modi signiWcandi were more fundamental than others. The essential
one deWned a word as a particular part of speech—noun or verb, for
example. Other accidental ones allotted to it such features as case, number,
tense, or mood. Complicated rules were worked out to determine which
words, with which modi signiWcandi, could combine together to make a well-
formed sentence.
Broadly speaking, it can be said that the study of modi signiWcandi was a
study of syntax, while the focus of semantics was the ratio signiWcandi,or
signifying relation conferred by the Wrst imposition. The speculative gram-
marians did, however, seek to Wnd a semantic element associated with the
modes of signifying. The sense of an expression is Wxed by the combination
of ratio and modi: this was called its ‘formal meaning’, its meaning in virtue of
language (virtus sermonis). In modern terminology we might call this its
lexical meaning, its meaning as determined by the dictionary.
In a context of actual use, however, an expression also has a reference
determined by its sense. Faced with the Latin sentence ‘Homo appropin-
quat’ we may be told that this consists of the nominative singular of the
masculine noun ‘homo’, meaning man, plus the third-person singular of
the verb ‘appropinqu o’, meaning approach. This information is given us by
the virtus sermonis: but we may ask, in a real-life context, which man is
approaching; and this fact opens up a new area of inquiry. The modist
logicians had various suggestions to oVer here, but they were not taken up
by later generations of thinkers. Instead, there was a revival of terminist
logic, which had developed the theory of suppositi on to deal precisely with
issues of the relationship between sense and reference.
Ockham’s Mental Language
One of the most important of the terminist logicians of the fourteenth
century was William Ockham. Ockham oVers a novel system: a terminist
13 See J. Pinborg, ‘Speculative Grammar’, in CHLMP 254–69.
LOGIC AND LANGUAGE
143
logic that is nominalist, not realist. All signs, Ockham maintained, repre-
sent individual things, because there are no such things in the world as
universals for them to represent. He oVers a series of metaphysical argu-
ments against the idea that a universal is a real common nature existing in
individuals. If individuals contained universals, then no individual could be
created out of nothing, for the universal part of it would be already in
existence. On the other hand, if God annihilated an individual, he would
destroy simultaneously all other individuals of the same species by wiping
out the common nature (OPh. 1. 15).
A universal is a singular thing, and is universal only by signiWcation,
being a single sign of many things. There are two kinds of universal:
natural and conventional. A natural universal is a thought in our mind
(intentio animae); conventional signs are universal by our voluntary decision,
being words coined to express these thoughts and to signify many things.
The signs in our mind are put together to make mental propositions in the
same way as spoken signs are put together to make a vocal proposition
(OPh. 1. 12).
Ockham regarded these mental concepts as forming a language system.
Besides the spoken, conventional, languages like English and Latin, all
human beings share a common, natural language. It is from this universal
This doodle in a
Cambridge MS is the
earliest known
representation of
Ockham.
LOGIC AND LANGUAGE
144
language that regional languages derive their signiWcance. The mental
language contains some, but not all, of the grammatical features studied
by the modists. Thus Mental contains nouns and verbs, but not pronouns
and particles. The nouns have cases and numbers, and the verbs have
voices and tenses, but there are not diVerent declensions of nouns and
conjugations of verbs as in Latin grammar. If two Latin expressions, or
two expressions in diVerent languages, are synonymous with each other,
then, according to Ockham, they will correspond to one, not two, elem-
ents of Mental. It follows that in Mental itself there is no such thing as
synonymy.
Other logicians in later ages have from time to time endeavoured to
construct ideal languages in which there is no ambiguity or redundancy.
Modern formal logics can be looked at as such idealizations of certain
fragments of natural language: the propositional connectives like ‘and’,
‘or’, and ‘if ’, the quantiWers like ‘all’ and ‘some’, and various expressions
concerned with tense and moods. Ockham deserves credit for being a
pioneer in pointing out the idealization that is involved in applying formal
logic to natural language, even if we may smile at his readiness to transfer
idiomatic features of medieval Latin into the universal language of the
mind.
It is one thing when a logician constructs an ideal language for a
particular purpose, as an object of comparison to draw attention to
features of natural languages that are ambiguous or invite confusion. It is
another matter when logicians—medieval or modern—maintain that
their ideal language is somehow already present in our use of natural
language, and contains the ultimate explanation of the meaningfulness of
the way we use words in everyday speech. If this was Ockham’s intention,
then his invention of Mental was futile, for it serves no such explanatory
purpose.
In the Wrst place, there is a problem about the nature of the mental
entities corresponding to spoken and written nouns. Ockham himself
seems to have worried about this, and to have changed his mind on the
topic at least once. Initially he identiWed the names of mental language
with mental images or representations. These were creations of the
mind—‘Wctions’ that serve as elements in mental propositions, going
proxy for the things they resembled. Fictions could be universal in the
sense of having an equal likeness to many diVerent things.
LOGIC AND LANGUAGE
145
What is the status of these Wctions? Ockham, at this stage, maintains that
they do not have real existence, but only what he calls ‘objective existence’,
that is to say, existence as an object of thought. There are Wctions, after all,
not only of things that really exist in the world, but also of things like
chimeras and goat-stags which are, in the ordinary modern sense, Wctional.
When we think a thought, there are two things to be distinguished: our ac t
of thinking, and what we think of, that is to say, the content or object of
our thought. It is the latter that is the Wction and that features as a term in
a mental proposition.
Later Ockham came to regard this distinction as spurious. There is no
need to postulate objects of thought: the only elements needed to support
mental language are the thoughts themselves. Unlike a chimera, my-
thinking-of-a-chimera is a real entity—a temporary quality of my soul,
an item in my psychological history. When mental names occur in mental
sentences, it is as elements in the thinking of the sentence. Ockham does
not seem to have made up his mind whether they were successive stages in
the thinking of the sentence, or a set of simultaneous thoughts, or a single
complex thought.
There is good reason for Ockham’s hesitation here, because the analogy
between speech and thought breaks down when we consider temporal
duration. Spoken words take time to utter, and one word comes out after
another. The case is the same with mental images of words, as when one
recites a poem to oneself in imagination. But thoughts are quite diVerent:
the whole content of a judgement must be present at once if a judgement
is to be made at all, and there can be no question of the temporal sequence
of the elements of a thought.14
However mental names are conceived, in Ockham’s view they all refer
to, supposit for, individual objects, since in reality there are no such things
as universals. These individual objects, however, may include individual
thoughts. Ockham’s nominalism means that he has to modify the theory
of supposition that we have seen in earlier logicians such as Peter of Spain.15
Ockham redeWnes the principal forms of supposition: simple supposition
and personal supposition.
Simple supposition had been deWned as a word standing for what it
signiWes; and this was taken to imply that in a sentence such as ‘Man is
14 See P. T. Geach, Mental Acts (London: Routledge & Kegan Paul, n.d.), 104–5.
15 See p. 130 above.
LOGIC AND LANGUAGE
146
mortal’ the subject ‘man’ stood for a universal. But for Ockham, simple
supposition occurs when a word stands for a mental entity, as in ‘man is a
species’, in which ‘man’ stands for a mental term, the only kind of
thing that can be a species. This is not a case of a word standing for
what it signiWes, for the term ‘man’ signiWes nothing other than individual
men.
In personal supposition it is indeed true that a term stands for what
it signi W es. In ‘Every man is an animal’ the word ‘man’ stands for what it
signiWes, because men are the very thing it signiWes—not something that is
common to them, but the very men themselves. But there can be personal
supposition even when a term is not standing for a thing in the world.
‘Personal supposition is where a term stands for what it signiWes, whether
that is an extra-mental reality, or a word , or a concept in the mind, or
something written, or whatever is imaginable’ (OPh. 1. 64).
Personal supposition is basic for Ockham, and it can apply to predicates
as well as subjects. A predicate signiWes, and supposits for, whatever it is
true of. Thus, if Peter and Paul and John are all the men there are, then
both in ‘Every man is mortal’ and in ‘Every Apostle is a man’ the word
‘man’ supposits for Peter, Paul, and John. This seems to mean that the Wrst
sentence is equivalent to ‘Peter and Paul and John are m ortal’ and the
second to ‘Every Apostle is either Peter or Paul or John’. A general term, in
other words, is equivalent to a list of proper names—a conjunctive list in
the Wrst case, and a disjunctive list in the second.
Truth and Inference in Ockham
Ockham uses the notion of supposition to deWne truth. A proposition like
‘Socrates is human’ is true if and only if the subject term ‘Socrates’ and the
predicate term stand for the same thing. This is sometimes called a two-
name theory of truth: an aYrmative categorical proposition is true if it puts
together, as subject and predicate, two names of the same thing. But
Ockham’s theory is a little more complicated than that, at least if we are
thinking of names as being proper names. As we have seen, for Ockham a
general term is not a proper name, but is equivalent to a list of proper
names; and the truth condition he lays down in terms of identity of
supposition amounts to the requirement that for an aYrmative categorical
LOGIC AND LANGUAGE
147
to be true one and the same proper name must occur in both the subject
list and the predicate list.
The simple two-name theory is easily shown to break down. If ‘Socrates is
a philosopher’ is true because Socrates can be called both ‘Socrates’ and
‘philosopher’, it is not easy to see how to explain the truth conditions of
‘Socrates isn’t a dog’. In order to know that ‘dog’ is not a name of Socrates,
we have to know what it is a name of: and there does not seem any answer
to the question ‘Which dog is it that Socrates isn’t?’ The more complicated
theory of Ockham does have an answer to this diYculty: the list corres-
ponding to ‘dog’ and the (one-item) list corresponding to ‘Socrates’ do not
have a common term. But it falls into a corresponding diYculty of its own.
If every general term is an abbreviation for a list of proper names, then every
proposition must be either necessarily true or necessarily false. ‘Socrates is
human’ surely is not simply a redundant identity statement. But that
is what it is if it means ‘Socrates is either Socrates or Plato or Aristotle’.16
Ockham devoted great attention to the logical relationships between
diVerent propositions: the theory of consequentiae, as it came to be called in
the fourteenth century. Earlier writers had used the word in the sense of
‘conditional proposition’. So understood, an example of a consequentia
would be
If Socrates is a man, Socrates is an animal,
with ‘Socrates is a man’ as the antecedent and ‘Socrates is an animal’ as the
consequent.
Consequentiae, so understood, could be true or false, and could be neces-
sary or contingent. Logicians were particularly interested in consequentiae
that were, like the example above, necessary truths. In such cases one can
construct a corresponding argument, namely,
Socrates is a man. Therefore, Socrates is an animal.
Here we have not one but two propositions, the antecedent here being a
premiss and the consequent being a conclusion. Arguments are not, like
propositions, true or false; they are good or bad, that is to say, valid or
invalid, depending on whether the conclusion does or does not follow
from the premisses.
16 See Kneale and Kneale, The Development of Logic, 268.
LOGIC AND LANGUAGE
148
Fourteenth-century treatises on consequentiae were concerned with sorting
out good from bad arguments, rather than with assigning truth-values to
the corresponding conditional propositions. Arguments could contain any
number of premisses: Aristotelian syllogisms, which contain only two
premisses, were just a single class of consequentiae. Premisses and conclusions
could be of various forms: they could include singular propositions, and
not only quantiWed propositions such as occurred in syllogisms.
Ockham begins by distinguishing ‘simple consequences’ from ‘conse-
quences as of now’. A simple consequence holds if the antecedent can
never be true without the consequent being true, e.g. ‘No animal is
running, therefore no man is running’. An as-of-now consequence holds
if the antecedent cannot now be true without the consequent being true,
even if at some other time that might be the case. An example would be
‘Every animal is running, therefore Socrates is running’, where, once
Socrates is dead, the antecedent can be true without the consequent
(OPh. III. 3. 1)
A second distinction that Ockham makes is between consequences
whose validity is internal (per medium intrinsecum) and those whose validity is
external (per medium extrinsecum). A consequence is valid externally if its
validity does not depend on the mea ning of any of the terms in the premiss
and conclusion. In such a case the consequence can be stated in schematic
form, using only variables: e.g. ‘If only As are Bs, then all Bs are As’.
A consequence is valid internally if its validity depends upon the meaning
of one of the terms: e.g. the validity of ‘Socrates is running, therefore a man
is running’ depends on the fact that Socrates is a man. There is no general
principle ‘If X is running, therefore an A is running’ (OPh. III. 3. 1).
Finally, Ockham distinguishes between material and formal conse-
quences. From the examples he gives it appears that he regards as formal
consequences both those that are externally valid and those that are
internally valid. In material consequences, on the other hand, the impossi-
bility of the antecedent’s being true without the consequent depends not
on any connection, external or internal, between the content of the
antecedent and the content of the consequent. It arises either from the
antecedent’s being necessarily false, or from the consequent’s being neces-
sarily true. Thus ‘If a man is an ass, then God does not exist’ and ‘If a
man is running, then God exists’, are both valid material consequences
(OPh. III. 3. 1).
LOGIC AND LANGUAGE
149
The Wrst of these is an instance of a general rule, ‘Anything whatever
follows from what is impossible’, and the second is an instance of ‘What is
necessary follows from anything whatever’. Ockham formulates a set of
such rules that apply to inference of very varied kinds. They include the
following six:
1. What is false does not follow from what is true.
2. What is true may follow from what is false.
3. Whatever follows from the consequent follows from the antecedent .
4. Whatever entails the antecedent entails the consequent.
5. The contingent does not follow from the necessary.
6. The impossible does not follow from the possible.
Many of Ockham’s rules derive from earlier philosophers, but he was the
Wrst to set them out systematically, and they were generally accepted by
later logicians.
Walter Burley and John Wyclif
In The Pure Art of Logic of Walter Burley the theory of consequences is given
even more prominence, and Aristotelian syllogistic is treated perfunct orily.
A very wide variety of inferences is brought under the rubric of ‘hypothet-
ical consequences’. The premisses of such inferences include not only
conditional sentences (containing ‘if . . . then’) but also conjunctive and
disjunctive sentences (with ‘and’ or ‘or’) and exclusive and exceptive
sentences (e.g. ‘Only Peter is running’ and ‘Everyone is running except
Peter’). An important class, studied also by Burley’s colleagues among the
Oxford Calculators, were sentences of the form ‘A begins to w’ and ‘A ceases
to w’.
Burley accepts Ockham’s distinctions between diVerent types of conse-
quence, and adds further subdivisions of his own. In all this, he is continu-
ing, sympathetically, work begun by Ockham. But when we turn from the
theory of consequences to the more old-fashioned topic of the properties
of terms, the picture is very diVerent. Burley rejects the nominalism that
Ockham had built into his logic, and restates the theor y of signiWcation and
supposition in a manner closer to its traditional realist form.
First, he rejects Ockham’s claim that a noun signiWes all the things to
which it applies.
LOGIC AND LANGUAGE
150
This noun ‘man’ has a primary signiWcation, and its primary signiWcation is not
Socrates or Plato. If that were so, someone hearing the word and knowing what it
signiWed would have a determinate and distinct thought of Socrates, which is false.
Therefore this noun ‘man’ does not have anything singular as its primary sign-
iWcation. So its primary signiWcation is something common, and that common
thing is the species. Whether this common thing is something outside the soul, or
is a concept in the soul, I do not much mind at this point. (PAL.7)
With ‘signiW cation’ thus deWned, Burley can restore the traditional deWni-
tion of simple supposition: a term stands for what it signiWes. The Wnal
sentence of the quoted paragraph leaves it open for his deWnition to
coincide in practice with Ockham’s deWnition of simple supposition,
namely that in simple supposition a term stood for a concept in the mind.
Burley not only defended, but also extended, the traditional theory of
supposition. As Ockham had done before him, he identiWed well-formed
sentences that were not covered by the types of personal supposition listed
by Peter of Spain and William Sherwood. One such sentence was ‘Every
man loves himself’: the classi W cation hitherto devised would not bring out
the fact that this entails ‘Socrates loves Socrates’. Burley said that in such a
sentence ‘himself ’ had a special form of personal supposition, half-way
between confused and distributive supposition, to which he gave a new and
complicated technical name. Another sentence which was ill served by the
traditional apparatus is ‘A horse has been promised to you’. In order to
distinguish between the case where you have been promised a particular
horse and the case where any old horse will be a fulWlment of the promise,
Walter had again to introduce new modes of supposition to assign to the
word ‘horse’ here.
As a critic of Ockham’s nominalism, Burley was soon outpaced by John
Wyclif, whose treatise On Universals is a sustained defence of realism. The key
to understanding universals, Wyclif believed, is a grasp of the nature of
predication. The most obvious form of predication is that in which subject
and predicate are linguistic items, parts of sentences. This is the most
discussed form of predication, and modern writers think there is no
other. In fact, Wyclif said, it is modelled on a di Verent kind of predication,
real predication, which is ‘being shared by or said of many things in
common’ (U 1. 35).
Real predication is not a relation between terms—like the relation
between ‘Banquo’ and ‘lives’ in ‘Banquo lives’—but a relation between
LOGIC AND LANGUAGE
151