20th World Congress of Philosophy Logo

Medieval Philosophy

Russell, Strawson, and William of Ockham

Sharon Kaye
Dalhousie University

bluered.gif (1041 bytes)

ABSTRACT: Realism and conventionalism generally establish the parameters of debate over universals. Do abstract terms in language refer to abstract things in the world? The realist answers yes, leaving us with an inflated ontology; the conventionalist answers no, leaving us with subjective categories. I want to defend nominalism — in its original medieval sense, as one possibility that aims to preserve objectivity while positing nothing more than concrete individuals in the world. First, I will present paradigmatic statements of realism and conventionalism as developed by Russell and Strawson. Then, I will present the nominalist alternative as developed by William of Ockham.

bluered.gif (1041 bytes)

Realism and conventionalism are commonly taken to be the primary contenders in the debate over universals. Does abstract language refer to abstract things in the world? The realist answers yes, leaving us with an inflated ontology, the conventionalist answers no, leaving us with subjective categories. In this paper I would like to defend a third possibility which aims to preserve objectivity without multiplying objects. It is nominalism, in the original, medieval sense of the wordor more specifically, in the Ockham sense of the word.

Willard Quine once remarked that "the nominalists of old . . . object to admitting abstract entities at all, even in the restrained sense of mind-made entities."(1) This is certainly true of Roscelin, the eleventh-century anti-realist who famously asserted that a universal is nothing but a flapping of the vocal chords. And Quine’s remark is true of Ockham as well, in so far as he asserted that a universal is nothing but a particular thought in the mind. Yet thoughts, even if particular, are not exactly concrete, and they do abstract, according to Ockham, in a way that Roscelin’s flapping vocal cords do not.

I won’t be able to defend Ockham’s nominalism by refuting all of the many versions of the competition one by one. What I propose to do instead is set it up in relation to the celebrated exchange between Bertrand Russell and P. F. Strawson. In this exchange, Russell and Strawson were trying to figure out how a sentence can be meaningful even when the thing the subject of the sentence refers to does not exist. Russell makes what I take to be the classic realist mistake; Strawson, the conventionalist. In what follows I will first explain Ockham’s alternative and then show why I think it compares favorably against these twentieth-century counterparts.

The key to Ockham’s nominalism, and what enabled him to go much further than his predecessors on the problem of universals, was supposition theory. Despite its well-known drawbacks as a formal logic,(2) supposition theory did have its advantages, one of which was the way that it clarified the disagreement between realists and antirealists over meaning and reference. We need to see what each of these linguistic concepts amounts to for Ockham in order to see how he explains abstract language.

First of all, what is meaning? Meaning is signification, of course, but there are two distinct kinds on Ockham’s view: conventional and natural.

Everyone agrees that written and spoken words signify by convention. This is evident in that there are a multitude of different human languages, all of which have very different words for things. Clearly, horses do not cause the spoken and written English word ‘horse’.

Nevertheless, Ockham thought that there is also mental language, and that it is not conventional but rather natural.(3) When perceived by a mind, a horse causes a concept of itself. This concept constitutes a mental word because the mind uses it as the subject of mental sentences. Whether you are English, Chinese, or an angel, a horse will signify itself to you in the same way. The meaning it gets across is what constitutes thinking of a horse and it is what we convey to others when we talk to them about horses. This is not to say, however, that the English word ‘horse’ signifies the concept or mental word for horse. That would leave us always talking about the things in our minds rather than the things in the world. Rather, the English word ‘horse’ becomes associated with the mental word for horse and thereby acquires the ability to signify the same thingi.e., real horses. To signify is to mean or to call to mind. First things in the world have this effect upon us, and then, with the help of conventionally agreed upon signs, we can cause each other to recall the effect.(4)

Although signifying begins in the perception of individuals, by generalizing our acquired concepts (through abstractive cognition) we are able to signify individuals which we have never perceived.(5) Let’s say, for example, that throughout your life you have had contact with a total of three horses. This is enough to cause you to understand the real similarity between them, and therefore to generate the concept for horse as a species or kind of thing. Despite the fact that this thought was triggered by Dobbin, Frenchy, and Quinn, it applies equally well to others. Of course, you have particular thoughts for each of the three horses which do not apply to others. Your thought of Frenchy, for instance, which has become associated with the name ‘Frenchy’, only signifies Frenchy. But the thought of horses in general which the three of them trigger in you does not only signify the three of them. This is evident in that when you talk with someone else about horses, he will know what you mean even though his experience was of a completely different set of horses. By saying the word ‘horse’ one can call to mind an abstract concept for all the horses there ever were, are, or will be.

So far we have been considering the word ‘horse’ all by itself. But, of course, human beings very rarely use words all by themselves. Rather, we employ them as elements, or what Ockham called terms in sentences.(6) And the term ‘horse’ does not necessarily refer to all horses past, present, and future. In the sentence

1) Horses were, are, and always will be four-legged

it does. This is evident in that we can restate the sentence so that it explicitly refers to the entire range of signification:

1a) All horses which ever were, are, or will be were, are, and will be four legged.

But in the sentence

2) Horses are more numerous now than they were in the paleolithic period

the reference of the subject term is not coextensive with its signification. This is evident in that the parallel restatement doesn’t make any sense:

2a) All horses which were, are, or will be were, are, or will be more numerous than they are now.

We can conclude from this that the thesis of natural signification is not enough to explain the role of a term in a sentence.

That’s what supposition theory is for. It tells us how the word is being restricted or modified as a term. Although the word ‘horse’ means or naturally signifies any and every horse past, present, or future, in sentence (2), the subject term refers to or supposits for presently existing horses only. The restriction can be even more extreme. In the sentence

Frenchy is a horse

the term ‘horse’ supposits for just one existing horse. We know this because the verb is present-tense and singular. At the same time, however, the term also conveys its broader natural significance. We know this because, if you said this sentence to Joe, he would know exactly what you mean (through his contact with other horses) even though he has never perceived the particular individual to which you are referring.(7)

So far so good: terms can supposit for all or just some of what they, as words, signify. And the medievals would label the sentences we have been considering cases of personal supposition. Things get confusing, however, when terms are used to supposit for things which they do not signify at all; i.e., when they supposit non-significatively. Suppose, for example, someone says the following sentence to you:

Horse has five letters.

What she has said is true. Nevertheless none of the things the word ‘horse’ naturally signifies (all of which are, were, or will be real animals) has five letters. What’s happened here is that she has used a spoken word to supposit for a written word. In so doing she makes it refer to something which it doesn’t mean.

Why has she done this? Well, she could have used the word which naturally signifies the thing she’s talking about. Just as horses trigger in us a general concept for horse, written words trigger in us a general concept for written word. So she might have said,

The written word which in English conventionally signifies those things (pointing to a group of horses) has five letters.

But of course that would be like describing Frenchy every time you want to refer to her instead of simply using her name. The truth is that ‘horse’ is the name of the word for horse. The medievals called cases in which spoken and written words are used to refer to themselves material supposition, and nowadays we call them cases of mention as opposed to use.

Although material supposition is a fairly obvious and uninteresting linguistic category in and of itself, it is important in so far as Ockham used it as a model for understanding the language of universals. Consider the following sentence:

A horse is a kind of animal.

Here the term ‘horse’ supposits for a kind of animal. Are kinds among the things that the word ‘horse’ naturally signifies or not? Clearly not. Only a horse can cause a concept for horse. And a kind is not a horse. On the other hand, a kind is not a written word or a spoken sound either. Yet the sentence seems true. What exactly does it mean?

Ockham’s answer is that this is a case in which a written/spoken word refers to a mental word. As we have already seen, horse-in-general is the concept the mind generates upon perceiving a number of horses. So just as ‘five-lettered’ is language about written language, and ‘Midwestern phoneme’ is language about spoken language, ‘kind of animal’ is language about mental language. Ockham designates a separate linguistic category for such cases:

Simple supposition occurs when a term supposits for a concept of the mind and is not functioning significatively. For example, in "Human being is a species", the term ‘human being’ supposits for a concept of the mind. For it is that concept which is the species. Nonetheless, the term ‘human being’ does not, properly speaking, signify that concept.(8)

The idea is that, although the word ‘human being’ signifies Albert, and Betty, and Conrad, and etc, the term ‘human being’ can supposit for (i.e., name or mention or refer to) the mental concept which signifies all these individuals while not suppositing for any of them at all.

It was medieval realists who originally named this type of supposition "simple", thinking that it was a case in which one is referring to the universal form of a thing alone without also referring to any individual instances of it.(9) They thought of the word ‘horse’ as primarily signifying the form of horseness. For them, therefore, simple supposition was a case of suppositing significatively, and personal supposition, in which one refers to instances, was not. Ockham turns this realist analysis upside down. If we understand cases of personal supposition as signifying in a primary way, then simple supposition can be understood as non-significative along the lines of material supposition. Voila nominalism: when you are talking about a universal, you are not referring to anything outside the mind.

Now that Plato and Aristotle have mostly gone out of style, you don’t find many philosophers openly insisting that words primarily signify mind-independent forms rather than things. There is, in fact, a dim reflection of the nominalist account of simple supposition when, in modern English, italics are used to mark universals which are being mentioned rather than used. Parallel to our use of the inverted comma to mark instances of mentioning, or material supposition, the italics convey that the universal itself, rather than the things the universal signifies, is being talked about. But without the nominalist thesis that the universal is a concept caused in the mind, philosophers quickly maroon themselves on the opposing banks of realism and conventionalism. I turn now to Russell and Strawson to illustrate.

Bertrand Russell’s early work in logic lead him to think about the relation between meaning and reference. He was troubled by sentences such as

1) The king of France is wise.

On the one hand, this sentence seems perfectly meaningful. On the other hand, since there is no king of France at present, it seems that the sentence isn’t about anything. Yet if it isn’t about anything, how can it be meaningful? After all, the sentence

2) The flang is drubble

isn’t meaningful precisely because it doesn’t refer to anything. Clearly then, sentence (1) must refer to something. What exactly is this sentence about?

Russell tries to find an answer in the meaning of the expression ‘The king of France.’ It is incorrect, he argues, to think of this expression as a name like ‘George IV,’ which is designated to pick out a particular individual in the world. Rather, it is properly thought of as a description. Just as we would understand the descriptions ‘French thing’, ‘monarchial thing’, and ‘male thing’, even if there were no things which fulfilled these descriptions, so also, we understand ‘king-of-France thing’, even if there is no thing which meets that description. To say that the king of France is wise is to say

There exists one and only one x such that x is king-of-France and x is wise.

Notice how ‘king of France’ has become a descriptive adjective on this construal, just like ‘wise’. And if we ask in virtue of what it is that we understand descriptions, the answer is: in virtue of understanding the corresponding universal. Russell writes

Awareness of universals is called conceiving, and a universal of which we are aware is called a concept. Not only are we aware of particular yellows, but if we have seen a sufficient number of yellows and have sufficient intelligence, we are aware of the universal yellow; this universal is the subject in such judgements as "yellow differs from blue" or "yellow resembles blue less than green does".(10)

Russell concludes that when someone asserts that the king of France is wise, you understand what she means because you are aware of the concept king-of-France. The sentence is meaningful because it is about something after all. It is about a universal.

The problem with Russell’s analysis is that it relies on an equivocationwhich is evident upon closer examination of the quoted passage. He wrote:

Not only are we aware of particular yellows, but if we have seen a sufficient number of yellows and have sufficient intelligence, we are aware of the universal yellow.

This is true in so far as he means that you can be aware of the thought in your mind which abstracts from perceived yellow things. But then he goes on to say that

this universal is the subject in such judgements as "yellow differs from blue" or "yellow resembles blue less than green does".

Once again, it’s true that the thought which abstracts from particular yellow things literally constitutes the mental subject of these sentences. But it makes these sentences true only in so far as it refers to something other than itself, i.e., to particular yellows in the world. This is to say that these sentences are cases of personal supposition. Suppose they were cases of simple supposition where the mental subject refers to itself, as in "Yellow (italicized) differs from blue (italicized)". Then they would still be true, but for the wrong reason. For the concept for yellow doesn’t differ colorwise from the concept for blue. On the contrary, concepts are not the kind of thing to have any color at all. Thus, it’s not the case that whoever asserts the sentence "Yellow differs from blue" is talking about universals.

And this is enough to reveal the flaw in Russell’s analysis of the assertion that the king of France is wise. True, king-of-France is a universal in so far as it is a common name for Louis I, II, III, IV, and a whole host of other individuals. But if this universal were referring to itself in that sentence, it would be false for the wrong reason. For a mental term is neither wise nor unwise. Russell sees that there is nothing outside the mind for the sentence to refer to, so he concludes that it must somehow refer to the universal. Why?because he thinks that if it doesn’t refer to anything it must be meaningless. But this is the classic realist mistake. If Ockham is correct, then former kings gave sufficient meaning to the concept king-of-France so that there is no need to find anything in the world right now to give it meaning. In fact, the whole reason why the sentence is false is that, although it is meaningful, it does not at present refer.

Peter Strawson was one of the first to call attention to the disastrous consequences of Russell’s theory of descriptions. His assessment is that Russell has

confused expressions with their use in a particular context; and so confused meaning with mentioning, with referring. If I talk about my handkerchief, I can, perhaps, produce the object I am referring to out of my pocket. I cannot produce the meaning of the expression ‘my handkerchief,’ out of my pocket.(11)

This would be Ockham’s sentiment exactly. When you say "This is my handkerchief" you supposit for the particular object in your pocket. But the word "handkerchief" or the expression "my handkerchief" cannot signify just that individual, since Joe knows exactly what you mean even though he has only had contact with other handkerchiefs.

Unfortunately, Strawson was just as untutored in medieval nominalist logic as Russell. And as a result he goes too far, saying:

So the question of whether a sentence or expression is significant or not has nothing whatever to do with the question of whether the sentence, uttered on a particular occasion, is, on that occasion, being used to make a true-or-false assertion or not, or of whether the expression is, on that occasion, being used to refer to, or mention, anything at all. . . the meaning of an expression is not the set of things or the single thing it may correctly be used to refer to: the meaning is the set of rules, habits, conventions for its use in referring.(12)

Strawson wants us to learn the lesson from Russell that significance has nothing to do with reference. But this is the classic conventionalist mistake. Surely if there were never an occasion on which a sentence involving the term ‘horse’ referred to a real horse then the word ‘horse’ would never have signified anything at all.

In his response to Strawson’s critique, Russell is just as adept as his opponent in pinpointing the problem. He writes,

So far as I can discover from the context, what [Strawson] objects to is the belief that there are words which are only significant because there is something that they mean, and if there were not this something, they should be empty noises, not words. For my part, I think that there must be such words if language is to have any relation to fact.(13)

Surely Russell is right. The word ‘horse’ would be nothing but a subjective musing like ‘flam’ and ‘drubble’ if there were no concept associated with it which was connected with some existing thing. But what he fails to see is that the reality in question that has given meaning to the word ‘horse’ is real horses, and once they have done so, there is no need for them go on existing. If all of the horses died out tomorrow, the word ‘horse’ would still signify the horses which used to exist. And it is precisely because there would not then be anything for the term ‘horse’ to supposit for that the sentence "Horses exist" would be meaningfully false.

Hence, Russell and Strawson volleyed back and forth, each knowing that the other had made a mistake, and each making his own in turn. But the problem of meaning without reference is one which at least some of the ‘nominalists of old’ had a handle on. It is a special case of the problem of universals which is solved by maintaining that there are no universals outside the mind but that universals do come to exist naturally within the mind. On Ockham’s view, these mental terms are not abstract objects, but they are acts of abstracting triggered by objects; and thus, they are objective. Language works because meaning and reference neither always nor never but rather sometimes coincide. My suggestion is that perhaps Ockham’s account of signification and supposition is exactly what we need to keep this all straight.

bluered.gif (1041 bytes)


(1) "On What There Is" in Word and Object, (Harper & Row, 1963), p. 15.

(2) See, for example, Michael Dummet, "Quantifiers" in Frege: Philosophy of Language (Duckworth, 1973), pp. 8-33.

(3) "Thus, it is said that the act of understanding by which I grasp men is a natural sign of men in the same way that weeping is a natural sign of grief. It is a natural sign such that it can stand for men in mental propositions in the same way that a spoken word can stand for things in spoken propositions" [Summa Logicae Pt 1, p. 81].

(4) This is a ‘causal theory of reference’ rather like the one advocated by Michael Devitt in Realism and Truth (Oxford, 1991).

(5) "It is not the case that the universal concept man precedes the intuitive cognition of a man. Rather, this is the process. First a man is apprehended by some particular sense, then that same man is apprehended by the intellect. Through this apprehension a general cognition arises [habetur] which is common to all men" [ibid., p. 557].

(6) "A term is nothing other than a component part of a proposition. . . . A written term is a part of a proposition inscribed in some body which the bodily eye sees or is able to see. A spoken term is a part of a proposition spoken by the mouth and able to be heard by the bodily ear. A conceptual term is an intention or an impression of the mind which naturally signifies or consignifies something and which is able to be part of a mental proposition and to supposit for the thing it signifies" [ibid., p. 7].

(7) It should be noted that in addition to categorematic terms (which are empirically acquired) one needs syncategorematic terms (logical connectives, adverbs, prepositions, etc) in order to be able to compose sentences. Ockham thought that these are hardwired in intelligent creatures: "We need to postulate mental nouns, verbs, adverbs, connectives, and prepositions. This is evident in that every spoken sentence corresponds to a mental sentence. If the parts of speech in a spoken sentence which are required for signifying are distinct elements of that sentence, then there should be distinct elements in the mental propositions which correspond to them. Spoken nouns, verbs, adverbs, connectives, and prepositions are necessary for all the different propositions and expressions making up spoken language. In fact, it would be impossible to express everything which we are able to express with all the parts of speech through nouns and verbs, for example, alone. Therefore, there must be distinct elements corresponding to all the parts of speech in mental sentences" [ibid., p. 14].

(8) Ibid., p. 196.

(9) See Paul Spade, "The Semantics of Terms" in The Cambridge History of Later Medieval Philosophy (Cambridge, 1982), pp. 188-196.

(10) "Knowledge by Acquaintance and Knowledge by Description" in Mysticism and Logic (Doubleday, 1957), p. 205.

(11) "On Referring" in Logico-Linguistic Papers (Methuen, 1971), p. 9.

(12) Ibid, p. 9-10.

(13) "Mr. Strawson on Referring" in Problems in the Philosophy of Language (Holt, 1969), p. 335.

bluered.gif (1041 bytes)


Back to the Top

20th World Congress of Philosophy Logo

Paideia logo design by Janet L. Olson.
All Rights Reserved


Back to the WCP Homepage