Discerning Between Local & Global Character Traits

In light of Fritz’s talk at the conference, I wanted to take a moment to think through the distinction that is fundamental to the situationist camp in social psychology. This is of course the distinction between local and global character traits. Roughly speaking, local character traits are a function of particular situations, whereas global character traits are invariant from situation to situation.

In order for the central claim of the situationist to be interesting, we ought to be able to clearly discern between character traits that are local and those that are global. But I find myself wondering just how easily these two might collapse into one another. The point can be approached from two directions. First, consider a putatively global character trait – for example, being cordial. It’s altogether too easy to conjure up a situation in which even the most cordial person grows peevish (e.g., subject them to torture). Thus it seems we were rash in calling this a global character trait to begin with, for we took certain aspects of the situation for granted. We are then forced to qualify the trait (thereby making it local) by augmenting it to being cordial-when-not-being-tortured. I suspect that similar stories can be told for any putatively global trait. Thus we might conclude that any global character trait can be rendered local by considering a wider and wider array of possible situations.

Or, in the other direction, consider a putatively local character trait – for example, being helpful-immediately-after-finding-money. The string of qualifications after ‘helpful’ is supposedly what makes this character trait local. But nevertheless, if a person possessed the local character trait of being helpful-immediately-after-finding-money, I submit that it would be globally (i.e. categorically, unrestrictedly) true of that person that they are helpful immediately after finding money. Here I am thinking of ‘it is globally true that…’ as a sentential operator, akin to the box and diamond of modal logic. In the same way that ☐ P implies ☐ ☐ P in most modal logics (which in turn implies that ☐ ☐ ☐ P, ad infinitum), I’m suggesting that for any local character trait a person has, it will be true of them that in any situation (i.e., globally) they have that local trait.

The natural move at this point would be to fall back on the idea that one character trait can be more local (or more global) than another, but that it might not make sense to speak of absolutely local or absolutely global character traits. I’m not sure if this makes the situationist’s claim too banal to be worth examining.

Degrees of Humean Responsibility?

Here’s a distinction that strikes me as interesting; unfortunately, any in-depth discussion of it is most likely going to be cut from my term paper. I’d love your thoughts on it.

I’m working with Paul Russell’s “naturalistic” interpretation of Humean moral responsibility. One of the most important principles that Russell attributes to Hume is the following: “A person or thinking being is held responsible if we regard her as an object of a moral sentiment” (64). This leaves open both of the following stances regarding responsibility:

(1)      Responsibility comes in degrees. When I disapprove strongly of a person (or, on perhaps a more plausible story, when a person’s action merits strong disapproval), I hold her very responsible. When I disapprove weakly of her, I hold her less responsible.

(2)  Responsibility is an all-or-nothing matter. As soon as a person arouses my moral sentiment (or behaves in a way that merits moral sentiment), even very weakly, I hold her responsible.

Now, within at least the reading that I’ve done, Hume tends to avoid the word responsible and its variants; he’s more comfortable with a more empirically accessible vocabulary (including words like praise, blame, love, and hate). So it seems to me that Hume, and Russell’s interpretation of Hume, would be compatible with either of these stances. Which one should a Humean (specifically, a naturalistic Humean) take on board?

Here’s a reason for the Humean to accept (1): it fits nicely with the rest of Hume’s thinking as outlined by Russell. If “regarding an agent as responsible is… a matter of feeling, not judgment” (64), it seems right to say that responsibility behaves roughly like our moral feelings do. And moral feeling certainly comes in degrees.

Here’s a reason that the Humean might want to accept (2): it seems to do less violence to our practice in attributing responsibility to people. I’d tend to call a person responsible for robbing a liquor store in just the same way that I’d call a person responsible for making a hurtful comment. But a good Humean might want to do away with this dimension of our practice.

Where should a naturalistic Humean side in this debate?

Hume, Necessitation, & Chance

A recurring topic in our discussions on Hume concerned his seeming refusal to leave any middle ground between necessitation and chance. There seemed to be some sort of a consensus that this conclusion of Hume’s is too strong. In other words, we don’t need complete necessitation of human action; at most, Hume’s arguments show that human behavior is to a large extent predictable and regular. But the fact that one’s character is a reliable indicator/predictor of one’s actions does not amount to the claim that one’s actions are fully determined by one’s character. Or so we might argue.

A defender of Hume could point out that he does in fact leave some space between total necessitation and pure chance/randomness. In particular, he seems to carve out this space as he leads into his definition of ‘liberty’:

“For what is meant by liberty, when applied to voluntary actions? We cannot surely mean that actions have so little connection with motives, inclinations, and circumstances, that one does not follow with a certain degree of uniformity from the other…” (Enquiry, Section VIII., paragraph 23).

Thus Hume thinks that any decent definition of ‘liberty’ ought to capture the fact that actions follow, with a certain degree of uniformity, from motives, inclinations, and circumstances – the subtext being that actions needn’t follow with perfect uniformity from motives, inclinations, and circumstances. As per our discussions in class, this seems like an apposite requirement to place on our definition of ‘liberty.’ But I suspect that it might contradict important claims that Hume makes elsewhere in Section VIII of the Enquiry. First, allowing uniformity to admit of degrees creates serious tension with the claim that every cause has a unique effect (ibid., paragraph 4). If every cause brings about a unique effect, then there wouldn’t simply be “a certain degree of uniformity” between cause and effect – there would be perfect uniformity. Second, allowing degrees of uniformity would seem to go hand in hand with admitting an element of chance into human actions – for what else but chance/randomness would account for the slack between motives/inclinations/circumstances and actions? The problem is that Hume assures us that chance “is universally allowed to have no existence” (ibid., paragraph 23).

I don’t doubt that Hume, or a defender of Hume, could wangle his/her way out of these tensions. I’m just not sure how the story would go. Any thoughts?

Hume and the Theory of Apparent Mental Causation

According to the theory of apparent mental causation proposed by Daniel M. Wegner, the late John Lindsley Professor of Psychology in Memory of William James at Harvard University, conscious thoughts do not cause actions. Based on a wide range of research into alien limb syndrome, schizophrenia, utilization behaviour, and the cognitive neuroscience of action, Wegner argues that the relevant conscious thoughts merely function as reliable previews because they (often) share with the action a common cause that is unconscious. On that basis the experience of conscious will is held to be an illusion in that it is a “feeling that comes and goes independent of any actual causal relationship between our thoughts and actions” (Wegner, 2002, x). Wegner claims (2002, 3) that his view that will is not a cause or force, but rather the feeling of such causing or forcing, can also be found in Hume (Treatise, 1739; pt. 3, sec. 1, par. 2):

… by the will, I mean nothing but the internal impression we feel and are conscious of, when we knowingly give rise to any new motion of our body, or new perception of our mind.

In a number of publications (e.g. Wegner & Wheatley, 1999; Wegner, 2002; Wegner, Fuller & Sparrow, 2003; Wegner, 2004; Wegner, Sparrow & Winerman, 2004) Wegner and colleagues argue that the experience of consciously willing an action occurs whenever the relevant conscious thought and action are understood by the agent to meet three criteria: they are compatible with each other (consistency); the thought precedes the action (priority); and no other potential cause conspicuously appears alongside the thought (exclusivity).

By manipulating the three variables experimenters have sought to induce the experience of will. For example, Wegner and Wheatley (1999) devised a Ouija-like cursor that was jointly moved by a subject and a confederate posing as another participant, around a screen displaying fifty small objects. Both were instructed to randomly move the mouse but to stop every 30 seconds or so. Subjects were also told that both participants would hear different soundtracks, containing music and words, over headphones in order to provide a mild distraction. In fact, however, the confederate was given instructions about which movements to make, and the word played to the subject was intended to prime thoughts about a given object prior to resting on it. Although the cursor movement was deceptively forced by the confederate, subjects – in accord with the priority principle – reported feeling that the movement was intentional with a strength inversely proportional to the time between the prime and the stop.

Since the experience of will is thus held to depend on inference based on external cues, Wegner further claims that Hume is “the early clear harbinger of this theory” (2004, 685) and that it “can be understood as an application of his [Hume’s] general analysis of the perception of causality” (2003, 67).

It is always gratifying to find a philosopher being discussed in contemporary science. Two questions naturally arise. First, is Wegner’s exposition of Hume accurate? For example, does Hume really take the will to just be a feeling when he defines it as an “internal impression we feel”? Second, do Wegner’s studies support Hume’s account of free will? On this point I have a provisional response to offer for further discussion.

In the Enquiry (1748; sec. 8, pt. 1, par. 21) Hume argues that some people mistakenly believe that they can perceive a necessary connection between cause and effect that goes beyond that of their constant conjunction. Assuming that such a perceived necessary connection would nonetheless ground constant conjunction, Wegner’s experimental results suggest that in the case of the will such a perception, or claim to have such a perception, is indeed mistaken.

However, Hume’s target is the person who claims to have such a perception in the case of external objects but not in the case of willing. Hume states that even in the former case claims to have such a perception are a pretence, and hence there is no reason to conclude that the determinations of the will are not necessary in the sense that the operations of matter are. Hume is concerned to argue that the only idea of necessity that we have is that of the constant conjunction of objects and the inference of the mind from one to the other. Wegner has merely provided empirical evidence that our conscious thoughts and our actions are not in fact constantly conjoined. Two further questions now arise. First, are Wegner’s theory and the evidence he offers on behalf of it compatible with Hume’s sceptical account of causation and causal inference? Since Wegner is doing empirical science rather than philosophy, I cannot see any reason to hold that he is committed to an account of necessity that goes beyond Hume’s. Second, Hume does not write of “conscious thoughts” but rather of “motives”, “volitions”, and “characters” that can be perceived to be constantly conjoined with actions. Do they always involve conscious thoughts, or can they subsume the postulated unconscious common causes in Wegner’s theory? Here, having little prior exposure to Hume and early modern philosophy, I am less sure.

Hume’s Definition of Necessity: AND or OR?

We’ve been treating Hume’s definition of necessity, more or less, as follows: necessity, properly understood, is the constant conjunction of objects AND the inference of the mind from one object to another. (Jorati, Handout 02/26 from Russell, SEP S.1).

Even Harris qualifies the idea of necessity in the following way: “When presented with one pair of usually conjoined events, our thoughts naturally proceed to the other; and, Hume suggests, this habit or ‘customary transition of the imagination’, is the best place to look for the source of the idea of necessary connection between bodies (Harris, 2005, p. 71).

The picture presented here is that two distinct elements are constitutive of necessity, for Hume: (1) the constant conjunction of events; (2) the inference from one object to the other. If I am not mistaken, the idea seems simple enough: we perceive a pair of objects that are usually conjoined and we naturally tend to infer one from the other. Call this AND-Necessity. Both (1) and (2) are constitutive of necessity.

Russell cites both A Treatise of Human Nature (TR hereafter) and An Enquiry Concerning Human Understanding (EnQ hereafter) as evidence for this view. Certainly, Hume is clearly committed to AND-Necessity in TR:

“Here then are two particulars, which we are to consider as essential to necessity, viz. the constant union and the inference of the mind; and wherever we discover these we must acknowledge a necessity” (TR, 2.3.1, par. 4)

But now consider (one of) Hume’s explications of necessity in EnQ:

“Necessity may be defined two ways, conformably to the two definitions of cause, of which it makes an essential part. It consists either in the conjunction of like objects, or in the inference of the understanding from one object to another. Now necessity, in both these senses, (which, indeed, are, at bottom, the same) has universally … ” (EnQ, p. 160, par. 27).

At first blush, this is quite different. Call it OR-Necessity. In this case, necessity is defined in one of two ways: it is either (1) or (2) (see above) rather than (1) and (2). The latter is true only in so far as (1) and (2) are “at bottom the same”. I’m not sure whether we should make a whole lot about this since Hume does, in other parts of EnQ, endorse AND-Necessity (p. 149-150, par. 5; p. 154, par. 16; p. 156-7, par. 21-2). In short, AND-Necessity implies that there are two distinct elements which are constitutive of necessity while OR-Necessity implies that ‘necessity’ can be used in two different senses, each corresponding to these two elements but where these two elements are ultimately the same thing.

I take it that Hume does not mean to be saying anything different when explicating OR-Necessity than he does when explicating AND-Necessity (and he does the latter consistently except for the passage just mentioned, as far as I know).  These, however, do seem importantly different to me and now I am unsure whether I have understood what Hume means by necessity if he means something that could be explicated in both these (apparently different) ways.

The Inference from Actions to Character

I have been puzzled by the nature of Hume’s ‘inference to character’ and was never quite satisfied that this was an obvious inductive inference. Giglio’s worries (below) convinced me that something more needs to be said about this issue. Russell noted that Hume is committed to the following: (1) for individuals to live in a society, we must be able to infer the actions of others from their character; (2) for people to regard others as responsible (for their actions), we must be able to infer their character from their actions. My worry is about (2) and I am not sure whether it would also extend to (1). (2) has never struck me as an inductive inference. It strikes me as an inference to the best explanation (IBE). That is, I infer from peoples’ actions to a motivation for those actions, or I infer from person’s X’s actions to a certain ‘character of X’ only in the sense that the latter best explains the former. ‘Characters’ or ‘motivations’ just don’t seem to be the kind of objects from which we reason inductively. They do not seem to be objects we perceive, let alone the kind of things that appear conjoined with actions. It seems we need to employ IBE in order to posit a character, or motivation, as the cause of an action. Of course, Hume would disagree:

“No union can be more constant and certain, than that of some actions with some motives and characters” (Treatise, p.260 par.12)

The worry, then, amounts to whether motives and characters are the rights sorts of objects such that it makes sense to say we perceive their constant union with actions. I’m not sure why I should think they are. Could someone help me see that I’m confused about this, and that we could construe this as a genuine inductive inference?

Hume & Abductive Inferences

In 1867, Charles Sanders Peirce first articulated abduction as the science of guessing or formulating hypotheses.  He placed it on a par with deduction and induction, claiming that all three of these patterns of inference are essential to scientific reasoning.  The first phase of scientific reasoning is the abductive phase.  Here an observation is made which seems to contradict some fragment of theory, and a hypothesis is put forward so as to explain this new, recalcitrant, bit of experience.  Next comes the deductive phase, in which consequences are deduced from the novel hypothesis.  Here we ask, “what would have to be true — what consequences would follow — if the hypothesis were true?”  And last is the inductive phase, in which we perform controlled experiments to ascertain the degree to which the results match up with the predictions formulated during phase two.

For an illustrative example, consider the discovery of the neutrino.  Around 1930, physicists were puzzled by the process of beta decay.  Carbon-14 had been observed to decay into Nitrogen-14, and in the process emit an electron.  However, if you do all the bookkeeping for this process, you will find that this leaves a certain amount of energy, momentum, and spin unaccounted for.  This was highly problematic, because the conservation of energy, momentum, and spin were taken to be fundamental and inviolable physical constraints.  Wolfgang Pauli then engaged in a piece of abductive reasoning: there must be another particle emitted which balances the equation, though it must be incredibly difficult to detect (extremely small mass, no electrical charge, etc.).  Lo and behold, roughly two decades later, the neutrino was finally detected with the help of some incredibly sensitive equipment.

So much for a primer on abduction; let’s turn to Hume.  It seems to me that Hume is by and large hostile toward abductive inferences.  To be sure, this logical vocabulary was not present during Hume’s time, but the pattern of reasoning itself is at least as old as Thales, who attempted to explain the presence of earthquakes by positing that the Earth rested afloat on water, and was periodically rocked by waves.  (A crude piece of abduction, but abduction nonetheless.)  Hume’s hostility emerges most clearly, I think, in arguing against the meaningfulness of our idea of necessary connection.  Take it as granted that, as a result of the copy-principle, we only have an idea of constant conjunction.  How, then, are we to explain the patterns and regularities that we observe on the billiard table?  Here is a simple and straightforward piece of abductive reasoning: collision is so unequivocally and ubiquitously conjoined with a transfer of motion because there really is genuine causation between bodies, though we are not in any way keyed into that causation via our impressions.  This is an explanation that Hume seems to pertinaciously ignore, opting instead to categorize our idea of necessary connexion as ultimately meaningless.  For this reason I say Hume is, in general, hostile toward abductive inferences.

However, he seems to betray this hostility in his attempt to account for the novelty that often arises from a situation that we antecedently took to be quite predictable.  For example: a clock ticks with regularity for days, and then spontaneously halts.  Here a reasonable explanation would proceed from positing “a grain of dust, which puts a stop to the whole movement” (Enquiry, VIII).  Or consider a piece of unexpected human behavior: “a person of an obliging disposition gives a peevish answer,” from which we surmise that “he has a toothache, or has not dined” (ibid.).  Generally speaking, Hume feels comfortable with hypothesizing “the secret operation of contrary causes” to explain irregular and novel events (ibid.).

To sum up, I am doubtful that Hume is entitled to the abductive inferences of the sort mentioned in the above paragraph.  That is because his eschewal of one blatantly obvious abductive inference (in the case of necessary connection mentioned above) leaves me thinking that he must in some way be hostile toward abductive inferences per se — perhaps because this style of inference is markedly less secure than deductive inference.  Maybe it is hasty of me to say that Hume is, in general, hostile toward abductive inferences.  But then what accounts for his eschewing it on the one hand and his embracing it on the other?  He can’t have his cake and eat it too.

Causation and Responsibility

I think there’s something powerful and interesting about the Humean point that there must be some causation happening, somewhere, for our attributions of responsibility to hold. I’m curious about how far back along a causal chain his point can safely be extended, so I’m going to break it down into small steps and ask you all to weigh in on their plausibility.

Consider the following claims.

  1. For you to be responsible for an action, it must have been caused by an intention of yours.
  2. For you to be responsible for an action, you must be responsible for the intention that caused it.
  3. For you to be responsible for an intention, it must have been caused by your psychological states (“internal character, passions, and affections”).
  4. For you to be responsible for an intention, you must be responsible for the psychological state(s) that caused it.

(1) and (2) seem uncontroversial and correct. I imagine that even incompatibilists will be able to accept them without any trouble. (3), I take it, is the central Humean insight. I certainly see its appeal. But I wonder whether an incompatibilist can just reject it by insisting that the connection between character and intention is something weaker than causation—maybe something like “being causally informed by.” Maybe in order for me to be responsible for my intention, that intention simply needs to be causally informed by my character; it can be caused by something else, like an agent-cause or a quantum indeterminacy. Of course, this sort of response requires some story fleshing out the difference between being causally informed by x and being caused by x.

Assume that (3) works, though. I think that Botterill is probably right when he claims (on p. 299) that Hume’s point can’t progress from (3) to (4). Kevin has a comment on a post below (“One More Link in the Chain”) where he explains how some compatibilists, like Nomy Arpaly in Unprincipled Virtue, hold the following thesis: the causal origin of a psychological state doesn’t affect whether I’m responsible for acting from that state. So not all thinkers (not even all compatibilists!) would necessarily see the force of this further step.

But there is an intuitive force to (4), and some compatibilists would be happy to accept it. Fischer and Ravizza, for instance, claim that we must “take responsibility for” the psychological mechanism that produces our actions before we can be held responsible for those actions. And I imagine that any libertarians who think that psychological states can cause actions would be fine with (4) as well.

So I have a bunch of questions. Are (1) and (2) as uncontroversial as I’m assuming? Would it be more defensible for an incompatibilist to reject (3), or to accept both (3) and (4)? And if, as I’m arguing, there’s a much stronger intuition behind (2) than (4), what does that say about where we locate responsibility?