[69a] ‘On the Material Character/Physical Object Paradigm of Art’, Art-Language, Vol. 2 No. 1, February, 1972, pp. 51-55.

On the Material-Character/Physical-Object Paradigm of Art

Terry Atkinson – Michael Baldwin

I

The potential evidence for a subsumptive judgement, which is admittedly inexhaustible, must be distinguished from its actually selected evidence, which is necessarily finite. Professor Richard Wollheim (1) has argued that ‘... for the mainstream of modern art, the appropriate theory is one that emphasizes the material character of art, a theory according to which a work is importantly, or significantly, and not just peripherally, a physical object’. If the statement is supposed to be retrospective, it’s not so bad: we have been stuck with reism etc. Assume, however (for the sake of allowing the statement a prospect of interest) that there is some predictive intent behind it.

We have to inquire what constitutes the continuing ‘mainstream of modern art’ and what the criteria of appropriateness are. From the former we may be able to extract some definition or elucidation of ‘mainstream’ which is predictively salient. Assume that Professor Wollheim asserts the following harmless proposition: ‘If I look at the art of the recent past, I will find that its mainstream is constituted by works which are importantly, or significantly, and not just peripherally, physical objects.’ From this it may be defended that the mainstream is not all rubbish: whether it is rubbish or not will presumably be determined by various (separate) criteria. Another position that might be taken up is one that asserts that there is an important relation between being a physical object (etc.) and not being rubbish. Which way round are things? Has ‘mainstream’ got an axiological slant?

It is fairly clear that Professor Wollheim is subscribing to the confused material-character/physical-object paradigm of art. A predictive assertion can furnish a prescriptive statement. Is Professor Wollheim asserting a law-like statement? And further, is he asserting that this law-like statement is a priori? And is he prepared to characterize the a priori in pragmatic terms? (By ‘a priori’ is meant ‘that... which we can maintain in the face of all experience, come what will’. This may not be very respectable as a notion of the a priori – it will do for now. (It can’t be more confusing than Professor Wollheim’s use of ‘physical object’ (etc.).)

With regard to actual scientific enquiry, it may be contended that there is no fixed boundary between ‘empirical laws’ and ‘categorical principles’. Empirical laws may function like categorical principles without necessarily foregoing their status as hypotheses open to empirical testing. This account rests on an acknowledgement that a conditional may be an analytic theorem. But the interpretive functions are exercised by synthetic conditionals expressing inductive generalizations as well. But the account (following Poincaré, Lenzen and Pap) admits, at the same time, that inductive generalizations are finally conventionalized and converted into analytic or definitive statements in virtue of extensive empirical confirmation.

What we may have here is a law-like statement which has been immunized against possible future experience insofar as, if an experience does not fit it, so much the worse for the experience – rather than ‘so much the worse for the law-like statement’. But one will have to take account of the self-evolving and self-correcting character of inquiry. The standards and norms of inquiry evolve out of inquiry itself and means of operation are formed (and become formative) with respect to future operations through successful performances of operations in the past; they do not call for extraneous conventions stipulated by methodologists.

It seems that ‘universal’ propositions are said to formulate modes of possible operations. In terms of their function, they are plans for guiding future experimental operations; genetically speaking, they formulate habits that have proved to be of instrumental value in the resolution of problematic situations. Thus we may observe that the conditional ‘if mainstream art, then physical object’ can easily be discerned in both its functional and genetic role.

As a caution, it may be pointed-out that what functions as a major premise in one language may function as a rule of inference in another. For example, any theorem of algebra, such as the binomial theorem, may be a rule of inference with respect to arithmetic, but it is conclusion and potential premise within algebra. It has been suggested that it is but a matter of stipulation whether the rules of inference (syntactical transformation rules) in physics are to consist only of analytic ‘L-rules’ (i.e. logical or mathematical principles) or are to include synthetic ‘P-rules’. In fact, as scientific language becomes formalized, what were (originally) conclusions and potential premises tend to acquire the status of rules of inference. The fact that generalization is employed as inference rule reflects confidence in its ‘reliability’. If an inferred proposition turns out to be false, the responsible offender will be suspected among the premises, and not until this suspicion has been demonstrated as mistaken will the validity of the rules of inference employed be questioned. If the distinctive mark of a universal proposition is its instrumentality as a rule of inference, it may be said that any generic proposition tends to become a universal one as it is increasingly confirmed.

One way of detailing this is to discuss the distinction between material implication and analytic implication. The process of converting empirical laws into definitional truths could be formally refiected in transition from material to analytic implication.

If it were the case that analytic implications were based upon Aristotelian ‘real’ definitions that are beyond possibility of revision, the implication ‘For every X, if X is S, then X is F would rule out irrevocably the possibility that there should be found an X which is S but not P’. From the above implication, and the fact that A is not P, we could infer without empirical investigation, that A is not S. But, in fact, our confidence in the empirical law ‘All S are P’, which led us to define S in terms of P and thus to treat the universal sentence as a ‘logical truth’, may have been premature. Finding an instance that appears as S but does not exhibit P, one may decide to abandon the analytic implication as an inapplicable rule instead of deciding that the problematic instance is not a real S.

Principia Mathematica is an ‘extensional’ logic. Modal distinctions can’t be expressed in an ‘extensional’ logic except metalinguistically; hence, the symbolism of Principia does not facilitate anyone’s making a distinction between contingent (synthetic) and analytically necessary ‘A-propositions’.

In accordance with the traditional ‘square of opposition’ (‘For every X, “if X is S, then X is P” implies “There is no X such that X is S and X is not P”’) is valid in the system of Whitehead and Russell by the very definition of ‘implication’. But consider the two A-propositions ‘All crows are black’ and ‘All squares are rectangular’. Either one would be symbolized in Principia with the help of the familiar ‘horseshoe’ symbol which expresses the logical constant ‘if... then’. Consequently, by the definition of the horseshoe symbol, ‘All crows are black’ is equivalent to the negative existential statement ‘There are not any non-black crows’, and so-on for ‘All squares are rectangular’. But clearly, the ‘there are not’ clause has different logical force in the two examples: in the former, it is asserted that such and such is not the case, in the latter, that such and such cannot be the case.

In a well-formed differentiated symbolic language, it should be possible to distinguish symbolically not only ‘material’ and ‘strict’ implication, but further different kinds of analyticity; and an analytic implication that is based on an empirically grounded definition (or, conversely, a conventionalized empirical law) should be distinguished from an analytic implication that is based on mathematical definition. Otherwise, if ‘being a triangle implies being a plane figure’ and ‘being a whale implies being a mammal’ are indiscriminately referred to as ‘universal’ in the sense of ‘being necessary by “definition of a conception”’ an important difference is neglected: for the latter is analytic relative to a definition which embodies an inductive generalization.

A ‘constant conjunction’ (Hume’s terminology) may be said to be expressed by a generic proposition of the form ‘When and where A, then and there B’. It is (may be) explained by the interposition of a middle term C; it may appear as a necessary connection expressed by the form ‘if A then B’. If the premises in terms of which it is explained have a high degree of generality, it is, as it were, ‘made the associate of’ a whole class of other empirical laws which follow from the same premises. Once an empirical law is incorporated into such a class of empirical laws which may be explained simultaneously by a single comprehensive theory, we feel a certain pragmatic necessity in abiding by its validity: there is usually some reluctance to abandon a theory which unifies a considerable body of empirical laws. But the unifying body of empirical laws is itself synthetic (in the sense that it contains at least one synthetic proposition) and, hence, the particular law(s) it explains are open to nullification by experience. Quite often, the conversion from de facto conjunction to de jure connection goes unnoticed through the mechanism of habitual hand-me-down usage. On the other hand, if all the explanation of laws (empirical laws) by theories amounted to was that – the substitution of one de facto conjunction for another – there would be no inherent advantage to be gained from the explanations. An interesting theory, as contradistinguished from an ad hoc hypothesis, transforms several ‘facts’ into ‘facts’ with content and range.

There is a possibility for an analysis which treats ‘universality’ as a functional property which A propositions may have, no matter whether they be logically analytic or synthetic. Whenever an hypothesis is in the form of an A-proposition and is to be experimentally tested through application to concrete instances, it is recommendable to formulate it with explicit reference to experimental operations. It will then assume the form of a conditional sentence whose antecedent formulates possible operations and whose consequent predicts the observable effects of these operations.

Dewey asserted that ‘the universal hypothetical states the relation between the operation and its consequences.. .; it follows from this that any empirical hypothesis can be constructed as a universal proposition only so long as its operational meaning can be made explicit, i.e., stated in such a way as to indicate its ‘method of verification’. In its operational formulation, an hypothesis will usually have the structure ‘P implies that if Q then R’, where ‘P’ refers to the property that is hypothetically predicated and whose presence is to be tested, ‘Q’ to the testing operation, and ‘R’ to the expected effects of the operation whose occurrence confirms (or verifies) the predication of P. If the hypothesis is expressed with the purpose of asserting an invariant relation, or a uniformity, the mention of Q will usually be omitted, and could hardly be said to formulate a ‘possible mode of operation’ and thus to be universal. If, however, it is expressed with the purpose of directing experimental observations, ‘Q’ will appear in its formulation, and it will be universal or a-priori in function.

II

One thing is to consider explicitly what are the consequences of the affirmation (not much more) of physical objects and the like. Presumably we should look at the gestalt involved. But ‘physical object’ still needs sorting-out.

We may not be interested in anyone’s chattering about electro-magnetic fields as ‘art’.

One of the things that does come out is that Professor Wollheim can underwrite the traditional concepts of a person. Associated with this is the Sylvester relation: ‘being aware of every fibre of your body in relation to the work’ – or something structurally similar and just as mawkish as that. Sylvester apparatus is associated with ‘physical objects’. Certainly the traditional concept of a person is dependent on ‘physical objects’ in the macroscopic perceptible sense.

We have said that if Professor Wollheim’s statement is retrospective, then it’s not too bad: perhaps it’s obliquely empirical. But wouldn’t it be better to look at past work in terms of the fact that ‘physical object’ positions may be noticed around it? It seems that if the work is to satisfy anything more than consumership (or Lockean-political) criteria, it has to get around furniture. This may involve its being able to take in furniture in a technical way.

Isn’t the ‘physical object’ paradigm argument a tacit autonomy thesis? It’s not hard to construe it that way, and you don’t have to get that far away from the ‘scientific’ horizon.

What we are dealing with as de facto mainstream may be immanently associated with the ‘physical object’ paradigm. But it doesn’t follow from this that you can extract a satisfactory art theory from this (partial) immanence, or that you can’t understand the physical-object paradigm without being ontologically committed to it.

A working principle of Professor Wilfred Sellars (in Science, Perception and Reality, London, 1963) is that ‘The assertion that the micro-entities of physical theory really exist goes hand in hand with the assertion that the micro-entities of the perceptible world do not really exist’. We may not want such a complete revision. One type of Lockean ‘idea’ is introduced in a context of causal explanation – and concerns effects which are produced in the senses of percipients by ‘physical objects’, but which are produced in such a way that effects are not construed epistemically. There is another type of ‘idea’ which is a distinction within the sensory context considered epistemically. There is a problem for the so-called ‘new realists’ (Sellars) which comes from the apparent identification of the two. According to the ‘new realist’ account, the Lockean/Berkeleyan reduction of macro-physical objects fails because of the inherent squalor of phenomenalism. To admit the epistemic aspects of the sensory models is to admit, at least initially, both macro-physical objects and persons.

The breakdown of existential generalization and of the substitutivity of identity is not a symptom that ‘free’ terms refer to funny entities. The breakdown is a direct consequence of the fact that in modal contexts individuals have to be considered as members of more than one state of affairs.

It may be argued that there is, in ordinary usage, construction for talking about works (even perceptual terms) that are not prima facie modal. What seems to be the case is that that you deal with ‘physical objects’ as essentially related to a proposition modality – as intensional in some non-trivial way. So direct object construction is very familiar – that’s all – and again, you have to take ‘that’ clauses seriously.

It may be asked whether the ‘data’ that someone might try and introduce by argument from incomplete ‘identification’ (in theory – or by theory) can serve any of the epistemological purposes which such ‘data’ (and these are art theory ‘data’) are supposed to satisfy. The point is that you do not need to be committed to ‘mainstream-art-as-data’ in an ontological way. It is a fact, however, that if the characteristic features of intensional entities, as they are often conceived of, are attributed to what might be called ‘data’, it may be possible to ascribe to such ‘data’ some features which would make them epistemologically suitable. Professor Quine (From a Logical Point of View, Cambridge, Mass. 1961, pp. 151-152) says that any two ways of characterizing one and the same intensional entity (in ordinary modal contexts) must be necessarily equivalent.

Perception is modal. A perceptually presented object is identified with a physical individual in many cases; in other cases, a physically presented object is identified with a ‘perceptual’ individual. Which way round, if any way round, does Professor Wollheim’s proposed art theory have it? It seems that an art theory can’t deny the complexities of propositional modalities.

Perhaps what would be more appropriate for an art theory which could work with past art as an objectively given supply of ways (functions) in which more than one contingency could be dealt with. They would be part of an ideology, not an ontology.



Note



(1) ‘The Work of Art as Object’, Studio International, December 1970. (Revised version of a lecture delivered at the Gardner Centre for the Arts, University of Sussex, November 4 1970.)