Empiricism and the Metatheory of the Social Sciences

Roy Bhaskar

• Publisher: Routledge
• Year: 2018
• eBook ISBN: 9781351048422
• Edition: 1
• Topic: Philosophy
• Subtopic: Philosophy History & Theory

A picture has indeed held modern Western philosophy captive, that of the universe as a vast machine whose iron laws are best understood as exceptionless empirical regularities which, as it were, determine the future before it happens. This fantastic conception commands the assent, not just of positivistically-minded naturalists but of all the great anti-naturalists who champion a very different view of human action as a domain of freedom 'that somehow cheats science'.

The most fundamental move in Roy Bhaskar's system of philosophy, the germ of everything that followed, was to reconceptualise the natural world in transcendental realist terms, 'turning Kant around using his own method'. On this account, the universe is characterized by deep structures, mechanisms and fields that generate the flux of phenomena, and is in open, creative and emergent process. This completely recasts the terms of the debate between naturalism and anti-naturalism by remedying its false grounds and shows how philosophy can be liberated from its anthropocentric/anthropomorphic prison and rendered consistent with the best insights of modern natural science. There is necessity in nature quite independent of humans, but in an open world causation is multiple and conjunctural, the actual course of the unfolding of being is highly contingent and the bases of human freedom can be understood scientifically.

Written as a DPhil thesis when Bhaskar was in his mid-twenties, Empiricism and the Metatheory of the Social Sciences brilliantly launches this reconceptualisation and explores its implications for social science in the course of carrying through the metatheoretical destruction of empiricism. It will be indispensable reading for anyone interested in the development of Bhaskar's thought, in transcendental realism, and in the critique of empiricism, more generally of the philosophical discourse of Western modernity.

1 On the conditions of empirical description

1.1 Speech acts

Speech action consists in the intentional utterance of sentences. Speech acts are performed in and by speech actions. The reason why we need distinct notions of act and action is because, while it is normally possible to characterise the same action in different ways, the same act can normally be performed in or by different actions. Acts, then, do not function as the proper names of actions. The notion of an action is, of course, just the notion of what is done on a particular occasion, whereas the notion of an act is the notion of some specific thing that was done in or by an action. Acts, that is to say, are always acts of a particular type, presupposing some definite characterisation of what is done as an action of a certain kind (with a certain purpose, role or effect). The reason why the notions are sometimes confused stems, I think, from the condition that it is not possible to say what is done on a particular occasion without characterising what is done on that occasion in some way. The category of action is, however, not alone in this respect. Thus it is not possible to talk about an event, i.e. what happened on a particular occasion, except as an event of a particular type, involving (or presupposing) some definite characterisation of the event. It does not follow from this, however, that reference to the event presupposes some unique characterisation. For, if it did, it would be impossible to redescribe (or even misdescribe) that event without self-contradiction.

Speech acts are, paradigmatically, acts of communication. They constitute perhaps the primary form of social exchange. Speech acts are not, of course, the only acts of communication. Thus, a raised fist, a kiss, a handshake may all be ‘ways of saying something’. Neither, of course, are they only acts of communication.

It is by now generally accepted that a knowledge of the sentence uttered (spoken or written) alone will not normally be sufficient for the characterisation of the speech action as a speech act of a specific type, e.g. an assertion or a guess, a promise or a command. In addition, some knowledge of the context in which the speech action occurs will normally be required. In particular, a good (precise) characterisation may require information about the preconditions of utterance, the speaker’s intentions, the success of his performance and more generally the response of his audience (including the reactions of the speaker himself). For example, the characterisation of an episode as ‘explanatory’ normally requires the presence of a puzzle, or at least a problem, even if – as is often the case – this has first to be induced (or established) by the explainer himself before his explanation can begin.

Given that speech action normally involves communication, it is in principle possible to characterise the same speech action in relation to (a) the intention of the speaker, or (b) the response of his audience, or (c) the success of his performance. Thus suppose A says to B ‘S is P’ [or ‘Do x’] and B comes to believe that S is P [or doesn’t do x] as a result of A’s performance. We could now characterise A’s performance by saying

(a) ‘A asserted that S is P’ [or ‘A asked B to do x’], or

(b) ‘A got B to believe that S is P’ [or ‘A annoyed B’], or

(c) ‘A succeeded in getting B to believe that S is P’ [or ‘A failed to get B to do x’].

For completeness it should be noticed that speech action is sometimes characterised in terms of its success (or efficiency) in achieving some purpose or objective, without our thereby implying (or being committed to the implication) that it was the agent’s intention to achieve it. Thus we might say of an episode, ‘Well, I don’t know what he set out to do, but he certainly succeeded in amusing her.’ This gives us, then, a fourth way in principle of characterising speech action.

The distinction between my (a) and (b) forms is similar to Austin’s well-known distinction between ‘illocutionary acts’, viz. acts that are done in uttering a sentence in a certain context (e.g. urging, advising), and ‘perlocutionary acts’, viz. acts that are done by uttering a sentence in a certain context (e.g. persuading, deceiving).¹ Clearly, as speech action involves communication, it will be possible to characterise any such action in either an illocutionary or a perlocutionary way. Thus ‘commanding someone’ and ‘getting someone to do something in virtue of one’s authority’ are not typically used to refer to distinct performances (actions) in the way in which ‘murdering’ and ‘playing poker’ are (though, of course, murder may involve playing poker and playing poker may result in murder). Rather, they are alternative characterisations of the same action; or better, characterisations of different parts, stages or aspects of the same action, so that we could also say that they are different acts performed in the course of that action.

The reason why it is possible to characterise speech action in either an illocutionary or a perlocutionary way is, of course, because it involves communication. This is a point overlooked by a recent writer (Alston) in arguing that illocutionary (unlike perlocutionary) acts cannot be performed without the use of language.² This is quite arbitrary, for a finger can summon, a frown reprimand and a tear plead (illocutionary characterisations), just as a look can frighten, a smile distract or a cough relieve the tension (perlocutionary characterisations). It is perhaps also worth mentioning that it is because speech action involves communication that two interconnected problems often thought (at least by philosophers of positivist persuasion) to pose special difficulties for the characterisation of action do not arise, at least in such a sharp form. In the first place, as, for a wide class of speech acts, the speech actor must make clear his intention to his audience if he is not to be denied the possibility of success, the ‘problem’ of identifying an agent’s intentions becomes (merely) a problem of knowing when such acts are occurring. And second, as, with the exception of soliloquy, (at least some of) the consequences of speech action are expressed in the behaviour of a different person (or persons), the ‘problem’ of identifying an action independently of its consequences is to that extent eased. It is for these reasons that we do not find it difficult to give, as participants and observers, illocutionary characterisations of speech behaviour in social life.

1.2 Meaning

‘What is the meaning of the word “x”?’ means roughly either

(1) ‘What word (or phrase, etc.) could I use instead of “x”?’ or

(2) ‘What is the characteristic role (or function) of the word “x”?’

These ‘use-substitutes’ of our original question provide us with what are, I think, the two central meanings (in the sense of synonym) of word-meaning: meaning as synonym (as use-substitute) and meaning as role.

Now, as it is very rarely, if ever, that two words are completely synonymous, i.e. substitutable ‘without loss of meaning’ in all their uses, (1) must normally be interpreted as

(1′) ‘What word could I use instead of “x” in this particular use (or range of uses) of the word “x”?’

This immediately raises the problem, of course, of how one defines a use of a word – as part of a sentence or a speech action? A dictionary typically illustrates the meaning of a difficult word for which there are no exact or clear synonyms by providing a sentence that is characteristic of a certain context of speech action. Sometimes, however, it may be necessary to specify this in detail, including not only the illocutionary characteristics but also (in some cases) the normal or standard perlocutionary effects of their use. This might be necessary, for example, in the case of words such as ‘democracy’, which tend to have in our society strong characteristic, often non-substitutable, perlocutionary effects. In such cases we could say, following Peirce, that the word had outgrown its function as a ‘symbol’ pure and simple and had developed an ‘indexical’ power that was its characteristic utterance force.³

The idea is sometimes mooted that it is possible to give a complete explication of word-meaning in terms of sameness of use. This depends crucially on the assumption that it would be possible to characterise the speech acts concerned in a sufficiently precise way without making use of the word (or an equally obscure one) in the characterisation of the speech acts necessary to elucidate its use. More basically, I think, this programme rests on a tacit conflation of the two senses of word-meaning distinguished above. For, if a person wants to know what a word means (in the sense of synonym), then he reaches for a dictionary, but the dictionary does not tell him, nor can it always be inferred from what it says, what its characteristic role or function is. On the other hand, if a person wants to know what a word means (in the sense of role), he is asking for the characteristic role that the word plays in language or the function it serves in social life. And a statement of its role or function will not (in general) be substitutable for the word itself; neither will a knowledge of its role necessarily allow us to infer its most appropriate synonyms. We could say that, whereas the elucidation of the first strand of word-meaning represents the programme of the lexicographer, the elucidation of the second represents that of the linguistic philosopher.

Conflation of these two senses of meaning could, with some justice, be called the central category mistake in the theory of meaning. It is typically committed by critics of some theory which is most reasonably construed as attempting a functional explication of meaning, and consists in showing that a description of the function of a word is not substitutable for the word itself. From this, of course, it follows, not that the attempted functional explication is incorrect, but merely that a description of the function of a word is not necessarily its use-substitute. There is no reason why this should be any more surprising than the fact that we cannot use a description of the function of, say, an electric drill for the same purposes as the drill itself.

I think that Searle commits this category mistake in arguing that some philosophers, e.g. Hare and Strawson, have been misled into identifying ‘meaning’ and ‘use’ by concentrating on simple indicative language.⁴ Thus he sees the fact that the words ‘good’ and ‘true’ cannot be replaced with the words ‘commend’ and ‘assent’ in all their uses as a refutation of the theories that the meanings of these words are to commend and give one’s assent, respectively. Now if we clearly distinguish between the two senses of meaning, we can, of course, allow that the characteristic functions of the words ‘good’ and ‘true’ are to commend and to give one’s assent, respectively, without supposing that these words can be replaced with the words ‘commend’ and ‘assent’ in all (or indeed any) of their uses.

Searle’s argument is of additional interest in that it turns on the claim that, while it is possible to substitute ‘I commend this car’ for ‘this is a good car’, we couldn’t substitute ‘I wonder if (or I don’t know whether) I commend this car’ for ‘I wonder if (or I don’t know whether) this is a good car’. This is odd (a) because ‘I commend this car’ is not, in fact, a good use-substitute for ‘this is a good car’; and (b) because, even if we were to take use-substitutability as the intended (or only) criterion of meaning, the second sentence would not provide a fair test of it. For, as ‘good’ is an adjective and ‘commend’ is a verb, only if the former is prefaced by a verb like ‘is’ or ‘does’ can it be used to perform the speech act of commending, whereas the use of the latter always consists in, or some remark on, the performance of such a speech act. A fair test of use-substitutability would necessitate taking an appropriate adjective, e.g. ‘commendable’, whereupon substitution now reveals that there is little to choose as far as substitutability with their respective sentences goes – and it does not go very far – between ‘I commend this car’ and ‘I wonder if this car is commendable’.

Use-substitutability and role are the two central strands in a web of meaning that extends from the meaning of words to sentences, speech acts, actions and even (derivatively) events. Thus to ask for the meaning of a sentence is often to ask either for a use-substitute (i.e. a sentence that one could use instead of it) or for its characteristic function or role in language. It is also sometimes to ask for the intention behind, or a justification for, the use of the sentence. Perhaps the most philosophically interesting sense of the meaning of a sentence is, however, the sense in which to ask for its meaning is to ask for the proposition being expressed. This sense is very closely connected with the use-substitutability sense in as much as there is no way of answering a request for the meaning (proposition) of a sentence other than by providing a use-substitute for it. Though, of course, this does not mean that the proposition is the use-substitute (to suppose which would be to confuse meaning and criteria of identity). There is more to be said on this in a moment. In the meantime we should record merely that (a) no speech act consists solely in the expression of a proposition, speech acts always consist in something being done (or attempted) with a proposition; and (b) not all speech acts involve the expression of a proposition, e.g. commands that express what we might call ‘directive’ meaning or oaths expressing exclusively ‘emotive’ (or ‘expletive’) meaning.

It should be noticed that in asking for the meaning of a perlocutionary speech act we may be (implicitly) asking for an illocutionary recharacterisation, e.g. ‘Why are you lying to me?’. More generally, in asking for the meaning of an action, however characterised, we may be asking for its (underlying) intention, a justification, or an account of its social function, role or consequences. We may even in some cases be asking for a casual explanation of it. And through the idea that reasons may in certain circumstances be (or at least function analogously to) causes, this provides a link with those contexts in which we talk of the meaning of an event, meaning by this its significance, cause or consequences.

1.3 Propositions and directives

Consider the sentence frames:

(A) … the Prime Minister is resigning

(B) … the Prime Minister ought to resign

It is clear that there is a difference between saying that the Prime Minister is resigning and saying that he ought to resign. And yet a moment’s reflection will show that it is equally possible to affirm, deny, suppose, wonder whether, and ask whether the Prime Minister ought to resign as it is to affirm, deny, suppose, wonder whether, and ask whether the Prime Minister is resigning. The fact that we can perform many of the same speech acts with sentences formed from the two different sentence frames (and as we shall see there is a sense in which the exceptions exemplify the rule) suggests that their difference cannot be located at the level of our normal characterisation of speech action (which is not to say that there is not some level that would bring out the difference). At the same time it is equally clear that there is a sense in which we could say that sentences formed from both frames were about the same thing, so that the phrase ‘the Prime Minister’s resignation’ expressed the idea of a topic common to both. The difference seems rather to lie in the way the topic is typically ‘reflected’ in the two cases. The use of sentences formed from the A frame expresses the idea of the topic [thought of or conceived] a...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Dedication
• Table of Contents
• List of figures
• List of tables
• Abbreviations and symbols
• Author’s acknowledgements
• Editor’s preface
• Editor’s acknowledgements
• Introduction
• 1 On the conditions of empirical description
• 2 Empiricist theories of the production of knowledge
• 3 Explanation in open systems
• Bibliography
• Index of ‘technical’ terms