Lipton on Inference to the Best Explanation

I have been having a quick look at (the late) Peter Lipton’s book Inference to the Best Explanation as preparation for revising a philosophy of science module I am due to teach in January. Given that he supervised my PhD rewrite one might have thought that I would have read his book as, at the very least, a courtesy. But it seems that I cannot, or cannot with any care as it seemed quite unfamiliar.

One key respect in which it proved unfamiliar is in a central aim. Lipton aims to use inference to the best explanation (IBE) to shed light on the kind of inference deployed in science. Rather than using inductive inferences to arrive at a theoretical view of the world and then, distinctly and subsequently, using explanation to shed light on particular phenomena, he proposes that the initial inferences can best be understood as being based on explanation. Science works by looking for explanations. The best potential explanation is likely to be true.

That general idea might be backed up / unpacked through any number of different analyses of explanation. Lipton himself – in what seemed to me to be the best aspect of the book – suggests that the best approach to explanation is contrastive. We do not simply explain a fact but rather a fact rather than a foil. Not just ‘why does Smith have paresis?’ but ‘why does Smith rather than Jones have it; or Smith when Jones does not?’. (The answer that Smith had untreated syphilis will be helpful if Jones did not, but not if he too did. The context matters to the answer sought: the relevant bit of the causal history.) Further, he makes a good case for the irreducibility of contrastive explanation to non-contrastive such as a deductive nomological model of explanation. It is not the case, for example, that to explain why P rather than Q is to explain the logical conjunction of P and not Q.

So with contrastive explanation in mind, Lipton argues that scientific inductive inferences can be understood as based on explanations. Elsewhere he gives a short summary of a clinical inference discussed at length in the book:

A good illustration of this is provided by Ignaz Semmelweis’s nineteenth-century investigation into the causes of childbed fever, an often fatal disease contracted by women who gave birth in the hospital where Semmelweis did his research. Semmelweis considered many possible explanations. Perhaps the fever was caused by `epidemic influences’ affecting the districts around the hospital, or perhaps it was caused by some condition in the hospital itself, such as overcrowding, poor diet, or rough treatment. What Semmelweis noticed, however, was that almost all of the women who contracted the fever were in one of the hospital’s two maternity wards, and this led him to ask the obvious contrastive question and then to rule out those hypotheses which, though logically compatible with his evidence, did not mark a difference between the wards. It also lead him to infer an explanation that would explain the contrast between the wards, namely that women were inadvertently being infected by medical students who went directly from performing autopsies to obstetrical examinations, but only examined women in the first ward. This hypothesis was confirmed by a further contrastive procedure, when Semmelweis had the medics disinfect their hands before entering the ward: the infection hypothesis was now seen also to explain not just why women in the first rather than in the second ward contracted childbed fever, but also why women in the first ward contracted the fever before but not after the regime of disinfection was introduced. This general pattern of argument, which seeks explanations that not only would account for a given effect, but also for particular contrasts between cases where the effect occurs and cases where it is absent, is very common in science, for example wherever use is made of controlled experiments. [Lipton 2000]

Still, if explanation is to shed light on our inductive / scientific practices, the notion of best explanation (or best potential explanation) had better not depend too closely on what seems inductively compelling. So best explanation had better not be most likely explanation at the risk of circularity. Instead, Lipton emphasises the potential initial gap between inductive inference and explanation by distinguishing the likeliest and the loveliest explanation. The likeliest explanation is the explanation that is ‘most warranted’. (An example of a likely but unlovely explanation is that opium puts people to sleep through its dormitive powers. Probably true but unhelpful.) The loveliest explanation is the explanation which provides the most understanding, or is the most explanatory. If IBE relies on the latter rather than the former, it can provide a substantial, rather than a merely trivial, analysis of inductive inference. It can come as a surprise that a lovely explanation lies at the core of inductive inference.

But the cost of emphasising the distinction – in order to preserve the substance of the analysis – is that it prompts the following worry, which Lipton calls Voltaire’s Objection. Why assume that loveliness and truth are correlated? Why assume that we live in the loveliest world?

Lipton’s response to this emphasises, inter alia, the fact that to be a potentially lovely explanation requires already meeting a number of constraints on adequacy. (He does not have a knock down reply to the worry. His aim is to suggest that his account is no worse off than any other when it comes to Humean worries about induction in general and he takes Voltaire’s Objection just to amount to that, in the end.) But the problem of this kind of reply, this direction of travel, is that it undermines the contrast which gave the analysis substance. So the question is then whether the two ‘vectors’ can be balanced: to have enough distinction to shed some light on inductive practices whilst maintaining enough connection to avoid Voltaire’s objection. After all if the analysis were to work there would have to be a connection and a gap: the paradox of analysis, again.

Lipton, P. (2000) ‘Inference to the Best Explanation’ in W.H. Newton-Smith (ed) A Companion to the Philosophy of Science, Oxford: Blackwell pp184-193
Lipton, P. (2004) Inference to the Best Explanation, London: Routledge

Further thoughts on the likeness argument

There’s been quite a bit of discussion of Neil Pickering’s critique of what he calls the ‘Likeness Argument’ amongst students planning essays in the UCLan philosophy of mental health programme which has helped me rethink my attitude towards it. In his original paper, Pickering concludes feistily thus:

The likeness argument dominates the dispute over the reality of mental illness. As a matter of fact, no one has yet produced a version of it which achieves what it aims and claims to be able to achieve, namely to answer the question whether or not conditions such as schizophrenia really are illnesses or not. Disputes continue on all sides about which likenesses and differences are to count, and what they are to count for. I have argued that the likeness argument cannot close these disputes, because two of the assumptions it relies upon to do so do not stand up to scrutiny. In particular, the idea that the features which should decide the issue (as the first assumption states), are describable independently of views about the issue (as the second assumption states), is false. They are not; and so the likeness argument must fail. [Pickering 2004: 253]

Why then, does the likeness argument fail to settle the matter? In his book The Metaphor of Mental Illness [Pickering 2006], Pickering argues that it depends on two assumptions both of which can be questioned. He says:

If the likeness argument is to resolve this dispute two things must, I think, be taken to be the case:
that there are features of human conditions such as schizophrenia, which decide what category, or kind, these conditions are a member of, and
that, with respect to the presence or absence of these features, a condition such as schizophrenia is describable independent of the category it is assigned to. [ibid: 17]

The first is a general condition derived from a view of how concepts apply to things. The suggestion is that concepts apply in virtue of conditions having objective features. This stands in contrast, for example, to a view where all such concept application depends on an imaginative human judgement. The second is a more specific assumption relevant to the debate about mental illness. It is that the ascription of features to conditions – putative illnesses – can be made independently of a top down decision as to the illness-status of those conditions.

Pickering suggests that both assumptions can be questioned. Criticism of the first – which he calls the ‘weak objection’ to the likeness argument – typically depends on pressing the role of human interests and values in the formation of human concepts. Nevertheless, as he goes on to concede, for the moment at least, there need be no incompatibility between acknowledging a role for interests and values in setting up a scheme of concepts and its autonomous application.

The ‘strong objection’ turns on questioning the second assumption. Pickering argues that the ascription of features to conditions – putative illnesses – depends on the overall category – illness or not – into which they are placed. The argument for this is piecemeal. In each of three cases – alcoholism, attention deficit hyperactivity disorder (ADHD), schizophrenia – he offers competing descriptions of their basic features manifesting first an assumption that they are illnesses and second that they are not. The behavioural features of alcoholism, for example, can equally be described in terms of moral weakness or of causally determined pathological behaviour. The same data can be equally well interpreted in the light of opposing top-down theories. Arguing that this is a general feature of such contested cases, Pickering concludes that the features themselves cannot be used to determine to which overall category the condition belongs. (He goes on to note that this claim also undermines the first assumption and thus his earlier concession was temporary.)

Now in the past, I thought along these lines:
This criticism is, however, not as successful as Pickering suggests. His central claim is that detectable and observable features of a condition, a putative illness, cannot be described without begging the question of the pathological status of that condition. This is not, however, a surprising claim. If the correct description of the features is taken to imply a pathological status then, trivially, it cannot be independent of the overall status. But even if it merely provides evidential support, Pickering has really only undermined a foundationalist version of the likeness argument, one in which observational data can be established independently of any broader theoretical perspectives. But most accounts of scientific theories now accept the theory dependence of data and an essential holism in theory testing. Most would reject a foundational approach to data, as Pickering himself later reports [ibid: 167]. Rejecting foundationalism in favour of scientific holism does not show that the likeness argument (or some version of the likeness argument) cannot work as part of a broader investigation of illness...

But now I am not so sure. Commenting on the suggestions made by two of our students, my colleague Gloria reported her own thought that the concession Pickering himself makes to a defence of the likeness argument against his own ‘weak objection’ is too generous.

Sticking with Pickering’s objection to the first assumption made by likeness argument theorists, both responses identify the qualm I have. In the case of the debate about the reality of mental illness, Ann makes the point that “There seems to be no reason why anyone should accept the similarities as being more significant than the differences,” and Rina follows this up with the observation that “the role of the observer/categorizer/diagnostician is not in the background” (and again that “human interpretation is definitely not in the background as Pickering argues).” The two points seem to me to be closely related: if we have no independent reason for accepting one set of similarities over another (e.g. anatomical lesions over reduced fertility), then the spotlight is going to turn on the person asserting a preference for one set of similarities in comparison with the dissimilarities. We want to know why he privileges just that set, and my suspicion is that the answer is likely to be circular: the likeness theorist probably privileges that set because he already thinks that mental disorder is an illness (Kendell) or he already thinks that mental disorder is a myth (Szasz). In other words, it seems to me as though these theorists have already decided how to conceptualise mental disorder, and then they set in search of similarities/dissimilarities that would serve the favoured conception. More generally, it looks like any likeness argument is going to depend on a favoured conception of mental disorder, rather than the other way around (i.e. that the likeness argument supports/justifies the favoured conception of mental disorder). This leads me to think that there is a vicious circularity underpinning the likeness argument, at Pickering’s Stage 1, and that this is more devastating to this strategy of arguing for the reality (or otherwise) of mental illness than Pickering realises.

There is surely something right about this. As a matter of fact, authors using a likeness argument strategy to argue either against or for the reality of mental illness have picked different criteria and produced different results. It is hard not to suspect that the criteria are picked for the results they produce. Given the dialectic, how else could the selection be disciplined so as not to be question-begging? My previous thought that Pickering was merely assuming and thus opposing a foundational version of the strategy seems to miss the point. How does that thought address this question (of how illness criteria could be selected without begging the central question of the debate for which they have been selected)?

I have today only two rather poor thoughts. One is that the debate concerns both the reality or not of mental illness as a whole but also the reality of mental illnesses. Whilst it is hard to see how a likeness argument could settle the former debate, it might help on the latter. The kind of question it might settle in the latter case concerns the pathological status of particular traits when, for example, the opposing view might be that the trait is a defence against illness, not part of an illness itself. In such a case, there might be sufficient gap between the general criteria of illness selected and their particular application to a novel case. (Of course, there’s no logical gap. But one’s reasons for adopting the general criteria need make no reference to / explicitly anticipate this particular case.)

Given the distinction between the general and the particular debate, I am not sure that a likeness theorist need even be embarrassed by no non-question begging answer to the first debate (about the reality of mental illness in general). But if so, the other thought picks up on my earlier mention of a ‘broader investigation of illness’. We might ask what we want a concept of illness for, what work it is supposed to do. From that inquiry would come a view of the general criteria and from that would follow a view of both mental illness and illnesses. That is, I think, how Jerry Wakefield sees his project: the idea that there is a failure of biological function plays a deep and partially hidden role in our use of the notion of illness. That connection forged, it can come as a surprise what particular results it has (via a likeness argument). But that connection was not made simply with a view to generating particular results.

PS: Original thoughts here. Yet further - more recent - thoughts here.

Pickering, N. (2004) ‘The likeness argument’ Philosophy, Psychiatry and Psychology 10: 243-54
Pickering, N. (2006) The Metaphor of Mental Illness, Oxford: Oxford University Press

So long ISCRI

Well that is just about it for the International School for Communities Rights and Inclusion, ISCRI. Last week I went to its wake: a pleasant enough meal in an empty restaurant in Preston (pictured) enmeshed in the coldest of winter snaps. By contrast with the 60 or so people who gathered for the first ‘away day’, 10 of us gathered at the end.

Gloria and I, and the philosophy of mental health, will move schools to the School of Nursing, to be renamed the School of Health, in January and I am looking forward to developing some of the themes of the philosophy of mental health for a philosophy for nursing. It might be a good place to develop my interest in the tacit underpinnings of clinical judgement.

Deriving a local urban social theory from Nietzsche

Yesterday, I listened to a talk by Grant Yocom of Oakland University called ‘The 'Last Man' in Detroit: Timely Revisions and New Targets for the Arrows of Longing’. It attempted to derive a kind of ‘local urban social theory’ from Nietzsche and apply it to an analysis of social phenomena in Detroit, a city in decline. Among other things, Grant argued that:

[Nietzsche’s] criticisms are aimed at the embodiments of mass-culture and the various forms value that these embodiments instantiate... [L]ooking at Nietzsche in terms of the concretized case study of post-industrial Detroit reveals that the meta-arguments and normative criticisms offered in his work only make sense when viewed in terms of a particular crisis. Specifically, on the fringe of Detroit we find a number of community organizations that instantiate embedded substantive and needs based forms of normative criticism still themselves beyond good and evil and emerging from the crisis-context in which we find them. [S]uch communities and organizations [are] fine examples of embodied Nietzsche-styled criticisms in action.

The community organizations in question have developed organically in response to a lack of proper supermarkets within the city limits by using abandoned lots for impromptu farms and Grant commended them as exemplifying a kind of creative life force in response to the real troubles the citizens faced. They exemplified the local evolution of ‘new values and new manners of living’ and contrasted with government imposed top down values and policies. Further, stressing their value (as he did) contrasted with a kind of middle class crisis-porn which pessimistically assumed that all was over for the city.

It was an interesting and engaging talk. But I couldn’t help thinking that there was something suspect about the project (of deriving a bespoke social theory from Nietzsche to fit Detroit) for this reason. Outside the admirable purity of philosophy as therapy, philosophy seems to me to have two key moves. It can either say ‘this is how things are because this is how they must be’, drawing on paradigmatically philosophical modal arguments. Or it can say, ‘this is how things should be’ in the more or less honourable tradition of offering normative suggestions, justified to a greater or less extent.

But in Grant’s case, it seemed highly unlikely that any philosophical arguments of the first sort could culminate in the inevitability of community urban farms. That would be too far a stretch for philosophical argument. But if it were the latter move, his dislike of imposed general top down values undermined a role for the very philosophical machinery he imported from elsewhere and applied to the case at hand. A non-philosophical description of the mere fact of such urban farms together with an invitation to agree they were admirable seemed all that was unproblematic.

If I followed his reply to my worry (and I fear I did not: my fault) it was to suggest that there was a very general structural account in accord with the first move (roughly: groups tend to respond well to tough ‘physiological’ challenges) and then a more or less local normative move to suggest that urban farms were indeed a good thing in the context. So rather than plumping for one move or the other, he adopted a delicate embrace of both. I am not sure how this would help, really. But perhaps in a Nietzschean further move, he suggested that the rival interpretation of the fate of Detroit (by Jerry Herron) that he had spent some time criticising might be equally descriptively correct. (That would, I assume, also have consequences for the normative stance to be locally adopted.) That was a nuance too far for me, however.

Priest on the principle of non-contradiction

An interesting short summary of his thinking on this is here. (This post is simply so I can find the link again!)