Our everyday judgements of the cause of events or happenings are intrinsically vital to every aspect of the theoretical and practical social sciences. Our question at hand is whether or not we can principally understand our judgements on the basis of the outcomes of previous similar experiences, or on our beliefs about the potentials of objects to be causal. Responsibility for our actions, and explaining the actions of unknown entities could be argued to be the very essence of human distinction from animals. The oldest of the enquiring disciplines of philosophy have wrangled with the notion of cause. For instance, White (1991) discusses how Aristotle’s understanding of the word ’cause’ was discretely different to our modern use of the term – where the ’cause’ is a larger concept than an immediate ‘reason’ for an event, but more towards a wider implication of a ‘purpose’ leading to a goal of development, or ‘telos’.
To sum up perhaps too simply, the enlightenment thinkers brought about a triumph of efficient causation and the abolition of final cause from scientific explanation (White, 1990). The conflicts of causal judgement concepts were played out between the figures of Mill and Hume. Mill, as Descartes and the classic philosophers before him, understood that:
“A statement of causation, is simply a statement of necessary and sufficient conditions.”
(Skorupski, 1998: 204)
This is important to remember when we do turn to our key contemporary theories on causal judgements, as the arguments inevitably encircle the extent to which Mill’s statement holds true. Lechalas said of Hume that:
“[For Hume], the cause, or better still, the set of causes, is the totality of conditions preceding the predication of a phenomenon…”
(Mehlberg, 1980: 79)
While Hume himself reflects that:
“‘Tis a general maxim of philosophy that whatever begins to exist, must have a cause of existence… but if we examine this maxim… on the contrary we shall find, that ’tis of a nature quite foreign to that species of conviction”
(Hume, 1880/1958: 79)
In other words, philosophers of this era were well aware of their challenges to convention of ‘general maxims’, and of the paradoxes of tackling causality.
This essay will attempt to review the prevalent psychological literature on the topic of causality but keeping Enlightenment rhetoric in mind, as ignoring the philosophical foundations of this long-contested debate would be foolhardy. Thus, we start with a common sense discussion of the double-edged question at hand: do we judge the cause of an event by the fact that something similar has happened before, or by the likelihood that the key objects involved have the tendency (‘power’) to be involved in the first place? These are over-simplifications on the subject of causality by Einhorn & Hogarth (1986) and White (1992) respectively.
The nature of all informed psychological literature is such that articles cannot disregard existing writing on a subject, so with each piece of research comes an increasingly-summarised review of the field. Einhorn & Hogarth (1986) begin by criticising theoretical views on causality:
“Psychology is not alone in presenting a fragmented view of causality… In philosophy, the definition and meaning of cause have been debated for centuries.”
(Einhorn & Hogarth, 1986: 3)
Their theory is that empirical regularities, or ‘cues-to-causality’ are systematically employed by humans for ‘assessing cause, both in science and everyday inference’. The relatively recent ideas of Mackie (1968, 1974; cited in Einhorn & Hogarth, 1986), are drawn upon in their construction of argument. Mackie argued that the degree to which a variable is a ‘difference-in-a-background’ determines its causal relevance, seen in three concerns over the conceptualisation of probable cause:
1) Characteristics of events that trigger causal reasoning
2) Role of background in distinguishing causes from conditions
3) How shifts in the causal field make alternative causes more or less salient
(Einhorn & Hogarth, 1986: 5)
Einhorn & Hogarth’s (1986) concepts to ideas of causality can be seen as distinct links to Mackie’s sentiments. Empirical evidence to back up any theory must look for differences between actual casual judgements of events between a theoretical control situation and a particular causal cue. Temporal procedurance, spatial contiguity, temporal contiguity & covariation have all been suggested as cues-to-causality (Einhorn & Hogarth, 1986). Tversky & Kahneman (1980: cited in Einhorn & Hogarth, 1986), demonstrated that when temporal cues ambiguous, the causal interpretation of data tends to receive disproportionate weight such that elementary rules of probability can be violated. In other words, in assessing judgements of cause, we use probabilities that re directly linked to temporal order. So the first of Einhorn & Hogarth’s cues-to-causality (temporal procedurence) appears to be satisfied.
Developmental psychology is one area that looks at causality in human behaviour. Siegler showed that in young children, temporal contiguity is a very strong cue for inferring causality, and that as we get older, we become less reliant on contiguity. Overall, when the temporal or the spatial contiguity is low, inferring causality becomes difficult. Einhorn & Hogarth (1986) give the example of the temporal gap between intercourse and birth, which if maintained is subject to out knowledge of biology and physiology. The 9 month gap in this case shows up their idea of intermediate causal models that fills in the blanks when causal judgements are not clear.
Classical conditioning as a paradigm makes a large use of causal inferences. In a typical single trial experiment pairing a taste of food with an induced gastrointestinal illness, Garcia (1981: cited in Einhorn & Hogarth, 1986), showed that in a sense rats are accurate judges of probable cause, despite a considerable temporal gap. We can see empirical examples in all aspects of psychology research, yet the most quotable and tellingly succinct statements are best seen in reviews of literature. Sedlak & Kurtz (1981) tentatively ranked three Humean principles on the basis of previous studies:
“Temporal antecedence supersedes spatial contiguity (Bullock & Gelman, 1979), and that temporal (and spatial) contiguity may sometimes outweigh the covariation factor in children’s judgement”
(Sedlak & Kurtz, 1981: 763)
This would appear to support Einhorn & Hogarth’s ideas that judgements of probable cause are affected by the a, b, c of alternatives (a discounting process for dealing with specific alternatives); backgrounds (Mackie’s concept of difference-in-a-background explaining variables’ effects); and cues-to-causality (probabilistic reasons we base our assumptions upon).
The cues-to-causality theory of causal judgement takes a large amount of literature, not least philosophical rhetoric – as its basis for empirical questioning. Einhorn & Hogarth (1986) present a holistic account of their beliefs by ensuring that they have algebraic expressions of causality to back up their ideas. While this essay takes a critical angle of the whole notion of judging probable cause, an analysis of the reviews of literature included as integral as the respective theories of Einhorn & Hogarth (1986) and White (1992) is somewhat redundant in the grand scheme of the question at hand.
Suitably, White (1992) takes a more theoretical and selective stance when answering when outlining his opposition to Humean regularity theory of causation and his own causal powers theory. This is not to say that he has somehow disregarded the past research, indeed his much-vaunted 1990 paper was a comprehensive review of the field, entitled ‘Ideas on causation in philosophy and psychology’. White inevitably reassures the contemporary attitudes to the cornerstone of the philosophical debate on causation – Humean regularity theory – but also looks at the seeming incongruencies of arguments arising from Enlightenment ideas. White cites Mill’s concepts on induction and scientific method as a continuation of Hume’s principles for identification of causes:
“Method of agreement (MA): If two or more instances of the phenomenon under investigation have only one circumstance in common, the circumstance in which alone all the instances agree, is the cause (or effect) of the given phenomenon.
Method of difference (MD): If an instance in which the phenomenon under investigation occurs, and an instance in which it does not occur, have every circumstance in common save one, that one occurring only in the former; the circumstance in which alone the two instances differ, is the effect, or the cause, or an indispensable part of the cause, of the phenomenon.”
(Skorupski, 1998: 125
Cross-overs with the basics of hypothesis tasting and the fear of false positives can be easily traced from such philosophical ideas. White’s strength is that he postulates a new model of causal judgements, without detracting from the empirical evidence behind the vast reviewing literature from philosophy and psychology disciplines.
The ‘causal powers’ theory takes as its basis two main premises on causal judgement: that we assess it on the powers or abilities of key objects involved in the event; and on the releasing powers under which a power yields to its effect. White asserts that empirical regularities are neither a necessary or sufficiency condition for causal inference.
Bhaskar (1975) shared the same sentiments on the non-necessity and sufficiency of empirical regularities for causal inference. His seminal work has this view on Humean ‘law’:
“It is this notion (that laws are constant conjunctions of events) that I intend to challenge… If atomistic events… constitute the world then you general knowledge to be possible, the relations between such events must be constant… (but it) they are not in general constant, then atomistic events cannot provide the only basis of ontology”
(Bhaskar, 1975: 64)
This argument against empiricism and Hume leads Bhaskar to suggest that if he is right in the above quote), the philosophical laws based on the empirical regularities in causal judgements ‘must be wrong’ (Bhaskar, 1975: 64).
White certainly picks up on Bhaskar, but gives his own syllogistic reasoning for the case against regularity (empirical regularity) theory. Take his example that under causal regularity theory, causal inference from a single observance should not be possible. White simply states that ‘this does not hold in fact’ (White, 1992: 162). Anecdotal evidence, such as when in human emotions we deduct the cause of say ‘anger’ as the person as being provoked, even though we’ve not necessarily seen this before – gives way to empirical studies. Michotte (1963) found that most subjects perceive a causal relation on their first exposure to a launching effect sequence. White goes on to acknowledge the reductionist analoguous approach which may be argued by regularity theorists, before giving a damning statement:
“Subjects would only be tempted to take an analogy with another kind of sequence which they believe you be a causal relation if they already had the idea that what they were perceiving might be a causal relation; and, under a regularity theory, they would have no such idea.”
A fundamental error of philosophy is referenced in White’s arguments against the necessity of empirical regularity for causal judgements, from Schopenhauer’s (and Hume’s) day-follows-night relationship. While the two always follow one another, regularity is not a back up for causal inference, so day does not cause night, but they are rather understood as discrete points on an eternal sequence.
Inevitably, White uses the opposing writing of Einhorn and Hogarth’s (1986) theory as a stimulus to argument. Einhorn & Hogarth say that a person striding twice has a high cues-to-causality that point to a causal relation. White accepts this to an extent, but correctly tells us that as judges we could not say that the second stride is caused by the first stride.
If people regard causes as necessary conditions, then Einhorn & Hogarth may have a claim that White is climbing down somewhat on his own statement, but he significantly says that even if regularity is neither necessary or sufficient for causal inference it doesn’t mean that regularity information is not used at all in causal inference. White concludes his article by giving a critique of Mill’s Method of Difference, in that it is too theoretical – when can two events where different outcomes ensue ever be exactly the same?
The robust theory of empirical regularity in causal judgement may have had its hey day in the time of Hume and Mill. No matter how we dress up the empirical evidence to ‘support’ the theories of either Einhorn & Hogarth or White, the self-juxtapositional paradox that ‘testing’ a hypothesis on the way we judge causality remains – insofar as no two situations are the same. The thoughts of one Quantum theorist should be borne in mind:
“It is obviously not the case that only physical events occur in time. So do psychological events”
(Mehlberg, 1980: 1)
White’s theory may make better logical sense than Einhorn and Hogarth’s cues-to-causality theory, but is this simply as the latter theory predates the former? Were White’s arguments incalculably changed by exposure to the theory of Einhorn & Hogarth? In their own logic judgements of cause must be a statement of necessary and sufficient conditions, yet even with preferences for one paradigm over another, human agency is fundamentally unchanged. The most ‘everyday’ causal judgements are made on the basis of what we know of the world, and if that embodies itself in ‘causal powers theory’, then psychology has itself a quotable source of philosophical debate.