Friday, October 15, 2010

Verisimilitude and Accommodation

As well asking whether a claim is true, we can also assess a claim by asking whether it is close to the truth. If we miss out on truth, verisimilitude seems like a good consolation prize.

There is an interesting question about the metaphysics of verisimilitude – what does the world have to be like for a claim to be close to true? This might be a matter of the similarity in relevant respects between this world and a world where it is true, for example – or maybe, while the talk of similarity is a helpful heuristic, or a metaphor with some truth in it, the sober story might need to be something else.

But in this post I do not want to talk about the metaphysics of verisimilitude so much as work that verisimilitude can do for metaphysicians.

Often metaphysical conclusions appear to go against things we normally say: there are many short-lived objects exactly where my stapler is, there are no tables and chairs, A=B yet there are features A has which B lacks, and so on. (I do not recommend saying all of these things at once.) In the terminology of Hawthorne and Michael, we can distinguish between compatibilist approaches to this apparent disagreement and incompatibilist approaches: the compatibilist holds that this conflict is only apparent, and that really what the metaphysicians says is consistent with what we normally say: maybe we normally tacitly restrict our quantifiers to ignore many things, or the metaphysicians speaks tenselessly and the folk normally speak in a tensed language, or the language of the ontology room is sufficiently different from the language of the street that “there are no tables and chairs in this room” in the ontologist’s mouth is consistent with what is expressed by “there are tables and chairs in this room” in a normal speaker’s mouth, even when the ontologist and the normal person are in the same room (I take it this last is e.g. van Inwagen’s view).

Dan Korman has been arguing in a number of places that these compatibilist strategies are usually inadequate (See for example here). On the other hand, incompatibilists seem to face a number of challenges, including saying what is good about many of the claims the folk take to be true but which are false according to the incompatibilist. (Maybe “there is a table in this room” is somehow wrong, but it’s not wrong in the way “there’s a hippopotamus in this room” would usually be.)

It seems to me that a metaphysician who claims that what ordinary claims say is very close to the truth in the usual cases has some of the advantages of both camps. She can (pretty much) agree with the ordinary claims, like the compatibalist, without hunting for a paraphrase or exotic semantic hypothesis. She can allow that the claims are strictly speaking false while having something a lot like truth, epistemically and otherwise, to attribute to the claims – they are close to the truth. That’s a pretty good status to have, one that is epistemically worth aiming for, one which we might think the demands of interpretive charity would be satisfied by, and so on. I’m not sure which side of the compatibilist/incompatibilist line we should put the verisimilitude option – Korman thinks it is a version of incompatibilism (at least he did when I asked him) – but whichever side of the line it falls on, it seems to me pretty close to the dividing line.

Compatibilists and Incompatibilists also can take a stand on what non-metaphysicians believe – to what extent is ordinary opinion consistent with their views, as opposed to what we ordinarily say? Here there is also a verisimilitude option – ordinary opinion is close to correct, or close to correct as far as it goes. There might be reasons for these to come apart – one may wish to think that people’s beliefs are often a little less committed than what they say, for example, in which case one might be tempted to think what is said is only close to true while what is believed might be entirely compatible with the truth. Though, as usual, it is often simplest to treat talk and thought together, and mark them both down as close to true.

The issue of what to say about ordinary beliefs, or orthodox theories, when one is a philosopher arises well outside the parts of metaphysics about which ordinary people might be thought to have views, of course. Metaethics is familiar with a variety of error theories, for example, and nearly every part of philosophy faces the challenge of saying what is good about some claim or intuition that is apparently rejected by a theory. Allowing that claims are close to true might provide a “comfortable” rather than radical kind of error theory about moral value, some intuitions about knowledge, or whatever else. (One application that has arisen around here at the moment is a way to sugar the pill of Alan Hajek’s thesis that most counterfactuals are false. If many of the false but apparently acceptable ones are close to true, while many of the false but unacceptable ones are not, that might help explain why we prefer the acceptable ones to the unacceptable ones.)

Of course, using verisimilitude to do philosophical work elsewhere does suggest that we should hope to clear up some of the philosophical puzzles about verisimilitude and how exactly it works. But a philosophical concept can be fruitfully used before all the puzzles with it are cleared up (see: every other philosophical concept which can be fruitfully used).


  1. Hi Daniel,

    I’d like to hear more about why the relevant beliefs should count as close to being true.

    Here are three ways for a false belief that p to be close to being true: (1) p is a generalization almost all of whose instances are true. (2) p attributes numerous properties to some object, and that object has most of those properties. (3) p attributes a single property F to some object, and that object has a property F* that’s a lot like F.

    Now consider the following belief, formed in a situation we'd ordinarily describe as one in which there are five cups and no other ordinary objects on the table:

    (*) There are exactly five things on the table.

    A belief with (*) as its content doesn’t seem to be close to true in any of these ways if universalism or nihilism is true. Of course, I don’t mean to suggest that 1-3 are the only ways of being close to true, but I wonder whether you can specify another way of being close to true that applies to (*).

    Also, even if the beliefs are close to true, I wonder whether that’s enough for an adequate account of folk belief. I would say that an adequate incompatibilist account needs to explain why the folk find it reasonable to believe these falsehoods in the first place. How does closeness to truth explain the apparent reasonableness of these beliefs? Is the idea that we have a (natural and blameless) tendency to confuse truths with close-to-truths?

  2. What about the thought that the content of (*) falls under (3) after all? Maybe the properties of having having exactly five things and the property of having five natural (enough) things --or place here the appropriate alternative restriction-- are a lot alike, in the relevant sense.

  3. Hi Dan K,

    I agree with Dan LdS that having five natural-ish things, or five familiar things, or something like that, is close in some important way to the claim that there are five things. It would be tough to say in advance all the respects in which closeness to the truth could be achieved - just as when we have a theory that is close to true (e.g. Newtonian mechanics) it is hard to say in what respects it is right and what respects it is wrong in advance of discovering an improved theory - even those who suspected that Newtonian mechanics was imperfect in the mid-nineteenth century would have been hard pressed to even imagine the ways it turned out to be not-quite-right thrown up by relativity and quantum mechanics (let alone the further ways it may fail to be quite true which future physics will reveal).

    Fortunately, somehow, we manage to assess theories as being close to true or not even in advance of being able to compare them with specific theories which are even closer to true - and so we could hope that the folk are tracking their theories as pretty good even if subtle investigations reveal there are things about our ordinary picture of the world that are not quite right.

    If I can appeal to good company, I think not just revisionary metaphysics faces the challenge of saying what was good about earlier ordinary belief - revisionary science does too. If the folk take there to be simultenaiety at a distance, or species-membership to be entirely intrinsic, or having the same heat in an object and its being the same temperature amount to the same thing, then a scientifically informed world view should be able to reflect on folk belief and say how some folk beliefs that endorse these mistakes are better than others. That's not an argument that the verisimilitude approach rather than some other approach is the way to evaluate goodness of ordinary beliefs, of course - there are other options around. But verisimilitude seems to me a good one to have on the table.