1. Freedom & Social Science: Some Arguments for Mild Compatibilism
2. Reasons & Causes: Some Objections to Radical Compatibilism
3. Behaviouristic Explanation: A Precis of the Cases For & Against
FREEDOM AND SOCIAL SCIENCE:
SOME ARGUMENTS FOR MILD COMPATIBILISM
According to Trigg (1985: 178-179), either any behavioural event depends on an antecedent condition, by which he means determinism, or else at least some behavioural events involve agency, an intrinsic independence from the reflexive physic of sociobiology. When he states that the possibility of human freedom 'haunts' the social sciences, he is restating the particular characterisation of the mind-body problem given by Ryle [1949] in his metaphor of the 'ghost in the machine', a concept derived from Descartes. Trigg restates the issue in terms of the incompatibilist view that if social science looks to behavioural causes in the sense of naturalistic necessary connections or inevitable extrinsic contingencies, then this programme is excluded to the extent that humans exercise intrinsic agency. He questions whether the realm of sociobiology and extrinsic contingencies can ever be the one true realm of social science. Based on incompatibilism, Trigg argues that naturalism is fatally flawed.
However, against incompatibilism are arrayed distinct and vying kinds of compatibilism: strong idealism and rationalism (matter is basically or emergently mentalistic and laws of thought govern causation, which may therefore be either deduced from prior conditions or determined by rational agency); strong materialism and behaviourism (mental phenomena are basically physical events, freedom simply refers to a type of complex data feedback in a human system which is in principle nevertheless explicable in terms of a total sociobiological physic); identity theories, perspectivism and monism (mind and body are identical, or are merely different perspectives of the same phenomena, or are aspects of one closed system: free choices from one perspective simply are the caused physical events of the other perspective); and Leibnitzian parallelism (mind and body are totally separate non-interacting categories which, whilst in some sense analogous, each world reflecting aspects of the other, do not require each other for explanation of behaviour as each is complete on its own).
It is important to observe that there are degrees to which determinism and freedom may be hoped to be reconciled, and various ways of saying that they cannot be. A strong compatibilism states that the same event may be both determined and free. Mild incompatibilism rejects this, but allows both determined and free events to occur in the same world, sometimes through the same agent. Strong incompatibilism rejects one possibility altogether. Examples of radical incompatibilism include that all events are determined and subject to a grand unified theory of extrinsic causation; or, secondly, all events are essentially intrinsic (instantaneous singularities, whether random or deliberate); or, thirdly, we are conscious of (at least) two worlds which are not causally linked but which are caused to appear to be in one-to-one correspondence by a necessity of thought, and in either of which determinism and freedom are incompatible (dualist incompatibilism).
The 'milder' incompatibilist version is roughly similar to the 'common sense' theory implied by ordinary language and our social institutions. In psychophysical interactionism, mind and body are separate worlds which nevertheless interact and cause events in each other, free action being that part of behaviour which originates in the mental world. Unlike Leibnizian parallelism, dualist interactionism denies thoroughgoing physical determinism on the rational ground that its own claim to theoretical validity is a product of rational argument supported by logical principles (Popper, 1977: 81), critical standards founded upon a fundamental 'philosophical' reality, rather than a mere contingent 'natural' utility, of reason (Lewis, 1976 [1947]: 18-19). If mental processes are entirely determined by non-rational processes, there would be no 'reason' to suppose that such a belief is true (Haldane, 1937 [1932]: 157). The imperialism of necessity seems to destroy itself by denying reason as a real insight, and therefore denying standards of theoretical adequacy which could recommend itself as a theory above others.
Herein is the problem for the social sciences, for if scientific explanation seeks to establish law-governed contingencies, how can human action flowing from mental processes involving 'reason' come within its domain of operations ? Papineau argues that determinism is compatible with freewill when the prior conditions of causality are "determination through the agent's beliefs and desires" (1978: 125-127). Similarly, Davis has argued that reasons explanations are both causal and deductive (1979: 89-91), the causal aspect being desire and belief. It is difficult to see how, by restricting 'reasons' to subjective interpretations in terms of goals and desires (Papineau, 1978: 78-83) rather than the more rigorous 'reasons' as analytic a priori propositions, accompanied by at least some synthetically perceived intuitive axioms and values, Papineau's reasons do not become extrinsic descriptive contingencies of the kind which exclude intrinsic agency.
A compatibilist solution with rationalist flavour might be the characterisation of conscious choices as tied to capacity for language, in which a margin of freedom is provided by the interaction of beliefs, meanings and symbols. This would surely only work with beliefs that were not merely descriptive of contingent states, but rather which were true 'revelations', direct and instantaneous knowledge of reason. The theory showing best fit with this dualism is Popper's species of interactionism, and especially the 'searchlight theory of consciousness' (Popper & Eccles, 1977: 472), in which rational choice is exercised as to which beliefs consciousness attends to, with consequent investment of mental energy ultimately translated into behavioural energy in the self's main organ, the brain. Popper suggests that the criteria for such choices may well be teleological, the organism choosing what kind of creature it wishes to become. Sperry has controversially suggested that this means that emergent properties and downward causation are implicated in human action which is free (Sperry, 1976: 9-19).
What is left of the explanatory programme of social science depends on the position adopted. Under 'libertarianism', in which human action is characteristically freely performed by agents who could have abstained and were not causally determined to act, it is argued that there can be no social science if reasons cause action since free acts would then be neither rational nor causal and therefore in principle inexplicable (D'Agostino, 1994: 31). Papineau argues that reasons cause actions even when the beliefs, desires and therefore the goals are totally irrational, since the means are always the agent's best reasons for action (Papineau, 1978: 83), and agents behave in terms of their best reasons. One of the problems here is whether the last statement is in fact a causal law. Propensities have been characterised as causal; but causal necessity or even constant conjunction of cause and effect cannot be demonstrated for a mere tendency. Such tendencies may exist in mathematical regions which in universal terms contain highly improbable random aggregates of events. Alternatively, they may be mere correlations implying causation in a third unknown prior event. Quantum mechanics has implied for the sciences the existence of objective probabilities for subatomic chaos. It is therefore highly problematic whether propensity is licensed as a causal law for non-chaotic events, still less for the incorporation into physical causation of something as different to chaotic randomness as rational thought.
The domain of social science can also be argued to be more than concerned with conscious, deliberate human action. It does in fact concern itself with the shadow of human action, its precursors, conditions, environments, unconscious and idiopathic parameters, individual and social impacts, and institutional products. None of these are free in the sense of libertarian doctrine, yet a libertarian who believed that human action was characteristically free might still without contradiction concern himself with them as the ground without which the form could not be properly understood or even adequately perceived. In this 'Gestalt' sense, mysticism may not be an impediment to social science, for two reasons. First, the way in which people choose their beliefs (acts of faith), direct consciousness to their visions, or choose meanings for their beliefs and utterances, may all be to some extent non-rational and non-causal, essentially free and intentional, but are available for naturalistic observation and study through the medium of language. Secondly, the concomitants, environment and social products which form the ground of 'mystical' freedom of action are available for rigorous causal analysis in the physicalist-determinist sense, and in Gestalt terms it cannot be said that a phenomenon does not include its ground.
Finally, Trigg's main thesis that 'the reality of human freedom restricts the scope of social science' (1985: 178) is not necessary if any one of the psychophysical accounts of reality, perspectivism, parallelism or interactionism, is true. In the last case, interactionism provides a very substantial and highly determinist programme for explaining the conditions and material requirements which restrict but never eliminate the intrinsic action of the 'ghost in the machine' (Popper, 1977: 104-109).
REASONS AND CAUSES:
SOME OBJECTIONS TO RADICAL COMPATIBILISM
Papineau claims that the way in which social developments can be freely negotiated whilst remaining determined by prior conditions is if reasons for action act as causes :
Compatibilists argue that an action's being determined is in itself no bar to our deeming it free…
[T] he compatibilist does not require absence of determination for free action. Instead what he asks for is determination of a certain kind, namely, determination through the agent's beliefs and desires (Papineau: 1978: 127).
The argument is that explanation requires that an event to be explained
is capable of causal explanation that rules out its non-occurence as impossible,
or (controversially) as at least highly and objectively improbable.
If rational grounds can be shown to operate on human action by a causal
law stating that agents behave, or have statistical propensity to behave,
rationally, then social science can have a domain based on rational explanation.
Papineau is able to accept an objection that people do not always behave
rationally :
This approach does, it is true, assume precisely that people are 'rational' in that they do what their beliefs indicate to be most likely to satisfy their desires. But it does not assume that those beliefs or those desires are themselves in any way rational (ibid.: 83).
The problem with unconscious behaviour and irrational beliefs for
Papineau's programme is that they interfere with the idea of 'freedom'.
It is precisely conscious deliberation and determination on the one side,
and possession of reason in the a priori sense on the other side, which
characterise free choice. It is on this basis that people who have
'lost their reason' or who are 'unconscious' are said to have 'diminished
responsibility'. It was not prudent for Papineau to attempt colonisation
of these regions with reason in order to argue that rational choices act
as causes. The remainder of human behaviour which is irrational could
be characterised as subject to extrinsic animal, ethological, behaviourist
and sociobiological determinism without denying a large margin of rationality
nor its causative character.
The philosophical objection to a scientific enterprise based on the compatibilist position of reasons as causes is that if reasons for action are also causes for action, then it is empirically impossible that such an action could not have occured, and therefore it could not have been free. Such identity between reasons and causes, it is argued, leaves as the only logically possible region of freewill the region which is neither rational nor causal, in other words a 'mystical' region unavailable to scientific explication by either programme (D'Agostino, 1994: 27-37). However, to deny that conduct was determined by the agent's motives is 'clearly to imply that there was no alternative' (Flew, 1979: 117). There are obviously two different categories of determination and causation being claimed, one an intrinsic singularity, the other an extrinsic chain of contingency.
Various solutions which allow a science of action without trivialising the meanings of causation or freedom have been tried by psychophysical theorists. One is the concept of emergent freewill and downward causation, which may be either monist and perspectivist (Sperry, 1976: 19), or dualist and interactionist (Popper & Eccles, 1977: 22-35, 87), although in earlier discourses Popper and others would seem to have refuted a social science based on 'emergence' as historicism :
If there is such a thing as growing human knowledge, then we cannot anticipate today what we shall know only tomorrow (Popper, 1961 [1957] : vi).
More mundanely, emergence as natural history
could only explicate social events if it were the case
first, that the formidable omission of the future does not conceal the point or meaning of the story, and, secondly, that we do actually possess history up to the present moment (Lewis, 1980 [1950] : 138).
A second solution to accommodation of freewill with social science
is simple perspectivism, either monist or parallelist, in which what seems
like free action is just this when looked at a priori, the same event being
causally determined in the absolute extrinsic sense when looked at a posteriori.
This would suggest a completely physicalist approach to social science
which acknowledges occasional correspondence with freewill but denies that
it makes any difference. The reality of freewill within a closed
mentalist system, it is argued, permits a claim to be able to ignore freewill
entirely for the purpose of scientific explanation (except possibly for
labelling) in favour of a non-interacting, closed physical-determinist
perspective of the same event. The problem for this programme is
that the two perspectives or 'worlds' plainly do interact, as implicitly
assumed by these theories themselves. Synthetic reason through mind
is a coloniser of the world of physical necessities, or no theoretical
knowledge would have practical issue. On the face of it, these are
not mere perspectives, but interacting, fundamentally different worlds,
linked by 'selves'.
A third psychophysical solution for a science of action combines rationalism with interaction between worlds of mental and physical phenomena (Popper, 1977: 36-50). Reasons and causes are different because they can be apprehended as different categories of reality, respectively transcendental and spatio-temporal, by minds which straddle both (ibid.). According to this theory, minds have a two-way causal relationship with matter, but relate only rationally as one-way receptors with a priori forms of reason (some synthetic, such as values or axioms; others analytic, such as logical grounds). Even if there are causal laws of propensity to act according to reason, the self has a margin of freedom to determine behaviour by its choice to vest energy or not in certain beliefs and desires. It may also be sometimes caused to do so sociobiologically and inevitably, and will always be conditioned by past actions and its environment, though not entirely determined. However, psychophysical theories stress that both the figure of free action and its ground of extrinsically determined behaviour are necessary objects of study to perceive the complexity of human action with any degree of understanding. Thus even if libertarianism does yield mysticism, the explanatory programme of social science would still be valid for understanding the concomitant physicalist and behavioural phenomena which surround, restrict and by complement delineate action.
BEHAVIOURISTIC EXPLANATION:
A PRECIS OF THE CASES FOR AND AGAINST
Skinner's concept of the environmental selection of behaviour by positive and negative reinforcement is basic to the behaviourist explanation of action in exclusive terms of analysis of the physic of extrinsically conditioned reflexes. Skinner's 'contingencies of reinforcement' in the behaviourist programme of explanation replace the vagaries of explanation in terms of 'motives' and 'reasons'. Looked at from outside, from the point of view of causation, prediction and explanation, Skinner thinks it is simply
a mistake to say that a man acted because of some desire or purpose, or in order to bring about some desired result (Skinner [1957] , in Lessnoff, 1974: 93).
His reasons for so thinking are rather perspectivist than simply
materialist. Mentalist explanation is characterised as an unwarranted
projection of the subjective perspective onto the objective phenomena of
action, which properly are only understood by physicalist analysis.
The analysis proposed is highly complex :
[W] e must take into account what the environment does to an organism not only before but after it responds. Behaviour is shaped and maintained by its consequences (Skinner, 1971: 24).
In any case, all behaviour is seen simply as the interaction of
genetic programming with environmental selection mechanisms, the latter
being vastly pre-eminent (including socialisation and social reinforcement)
in shaping behaviour by means of its 'schedule of reward and punishment'
(D'Agostino, 1994: 59).
It is tempting to criticise Skinner's radical behaviourism on the ground of its 'imperialistic' tendency to colonise areas which are more plausibly open to nativists on the one hand or rationalists on the other. Whilst implausibility alone is a matter for empirical rather than theoretical analysis, where a behavioural explanation postulates covert mechanisms of rationalisation which are in principle unfalsifiable (Popper, 1977: 442), the programme does seem to have the same major limitations as such unobservables as the Freudian unconscious or Marxist false consciousness. Yet as in thoroughgoing rationalist frameworks, equally radical naturalist postulates may be infered by systematic study of the correlatives of action, confirming indirectly by a 'patterned relation' between the rationalisation phenomena 'a place in a larger pattern of behaviour' (D'Agostino, 1994: 48). Some strident criticism of Skinner, perhaps due to his own strident manner, fails to notice that his programme amounts to a research focus and perspective that claims completeness within its own parameters. Despite a distaste expressed for other approaches, Skinner does not logically exclude simultaneous programmes of other (equally valid) foci or perspectives.
It is, in other words, possible to look at human beings from [minimally] two different points of
view - as matter, and as agents (Lessnoff, 1974: 94-95).
Chomsky offers two reasons to doubt the completeness of the
behaviourist account. First,
by rejecting the study of postulated inner states, Skinner reveals his hostility not only to 'the nature of scientific enquiry' but even to common engineering practice (1973: 104-110).
This criticism is perhaps unfair, as Skinner is at pains to emphasise
that both the indirect effects of the environment after the organism responds,
and the way in which behaviour is maintained by its own consequences, are
parts of behaviourist explanation (1971: 24). Chomsky's second charge
appears more pertinent :
[I] t is far simpler to assume that people are capable of a degree of choice in their use of language, and, by extension, in their social behaviour, than it is to speculate about unidentifiable stimuli determining every apparent choice (Chomsky, paraphrased in Mennell, 1974: 9-13).
Social behaviour is generally so complex that, particularly and
characteristically where mediated by linguistic behaviour, it is argued
that behavioural explanation is impractical compared with explanation that
assumes for practical purposes that people have goals and make choices
(ibid.). As a working hypothesis, the latter
by no means precludes a causal account… past choices become the causes of our future actions…
[P] eople's characters and ways of life become settled and established; therefore they are predictable, not as certainties but as probabilities (Mennell, ibid.: 9-13).
The behaviourist programme is therefore impracticable
for at least some phenomena, and therefore cannot provide a complete
account. In short, in some explanatory regions, behaviourism is nullified
by Occam's razor.
In addition, there are some human capacities which, it is claimed, could not in principle be explained behaviouristically. These include reasoning and valuing (Lewis, 1960 [1947] : 16-28; 38-42); creating and philosophising (Moore, 1903, in Flew, 1979: 221; cf. also Popper & Eccles, 1977: 16-17, 119, 458, 566); consciousness and self-consciousness (Sperry, 1976: 9-19; Popper, 1977: 61); first person avowals of inner sensations (Wittgenstein, 1953, in Flew, 1979: 37); and moral reasoning (Lewis, 1978 [1943] : 34-48). There is a problem in any totalitarian naturalism that such basic and primary realities become invisible -
To 'see through' all things is the same as not to see (Lewis, op.cit.: 48).
©1994 September- C.A. Newman
Other Academic Extracts, 1982-2000