Unconscious Inference

I received internalism about justification in my mother's milk, so to speak. Most of my earliest philosophical teachers were internalists in this way, and I have usually thought that this is about right when it comes to justification, despite my flirtations with proper-functionalism.

Recently, I have been more latitudinarian in my internalism - favoring what's been termed "mentalism" by Conee and Feldman in Evidentialism. Mentalism is a kind of bare internalism, it holds that what's justified for me to believe supervenes on my mental states. Thus, a brain-in-a-vat can have all the same justified beliefs that I have if he has all of the same mental states I do. That sounds about right to me (pace externalism about content worries). The upside of this form of internalism is this: my grounding for certain beliefs may not be easily accessible to me in the way that classical internalism envisions.

Philosophers often think (or it seems so) of the thinker as an intellectual, and perhaps even a caricature of an intellectual: a person who comes to all of their beliefs through conscious deliberation. But really, if I survey my own mind, and others seem to avow this, most of my beliefs are not formed this way: I have flashes of insight about theoretical matters when I'm working on something else (maybe after awaking from sleep, in the shower, while I'm working on an unrelated topic), I seem to make inductive inferences without going through any sort of explicit process (I see the groceries on the floor and immediately believe that my wife is home), I act in a way that shows that I have some sort of pro-attitude toward some proposition without it ever crossing my mind that I do (especially in skilled activity - like driving - I evade a road hazard without even thinking about it), etc. 

The basing relation with mentalism points the way toward explaining how we have justified belief (or some other doxastic/quasi-doxastic state) in these cases. Various mental processes, those I don't have direct control over dispose me to form the correct attitude from my evidence without the slightest conscious thought about it. These seem to be operating when I'm thinking about other things, when I'm asleep, when I'm driving, and so on. Often the only indication of these processes are when they seem to 'pop' into my conscious mental landscape, but even then, like in the case of driving, I only seem to be aware of them from the way my mind has directed my behavior.

Anyway, this seems interesting and bears more thought.

Content Externalism

(So, this is musing at this point... I'd love to be corrected by those who understand this issue better, especially if I'm making any simple mistakes due to my ignorance in this area. But blogs are good for this - exploring and figuring out when one has something seriously in error.)

I'm currently reading through Timothy Williamson's Knowledge and Its Limits[1] and one of the crucial issues in the book is Williamson's adoption of an "externalist" theory of mental content. I wish to muse a bit on such theories, since I am undecided about what the correct attitude toward externalism of this sort ought to be. I think Williamson is right in at least so far as he sees that the adoption of externalism about content will require some serious rethinking of traditional epistemology. 

Externalism has been typically argued for via "Twin Earth"-style thought experiments. I'll adapt for my own use one given by Tyler Burge. [2] Al is a fairly resident of Earth. Al has a variety of beliefs about aluminum. "He thinks that aluminum is a light, flexible metal, that many sailboat masts are made out of aluminum, that someone across the street recently bought an aluminum canoe."[3] Suppose that there is a duplicate of Al living on Twin Earth, a planet in every way like our own with one important exception, what plays the functional role of aluminum on Twin Earth. When twin-Al (call him 'Twal') runs his hand over twin-aluminum (call it 'twaluminum') he has just the same sorts of tactile stimuli as when Al runs his hand over aluminum. It weighs the same in each's hand, it reflects light in just the same way. Twal also calls twaluminum 'aluminum'. So, it would seem when Twal thinks of twaluminium, it's hard to see how his internal concept of twaluminum differs from Al's concept of aluminum. This is made more explicit on supervenience theses. If mental states, including intentional propositional attitudes, supervene on brains, then surely the mental state that Twal and Al are in are precisely the same. But if Al asserts, "This is a fine aluminum canoe" and Twal asserts, "This is a fine aluminum canoe" they surely are saying something different. Twal has never encountered aluminum and Al has never encountered twaluminum. But additionally, how could they be saying something different? If their mental act of asserting some proposition with those words supervenes on precisely the same (ex hypothesi) type of brain state, then surely they are making the same kind of assertative (sp?) act.

I think that it's intuitive, even with the rejection of a supervenience thesis about mind, that this problem occurs. I think that one of the central issues in these externalist thought experiments is that Al's and Twal's experiences underdetermine what it is that their intentional states are of. Suppose I don't have any beliefs about what the microstructure of aluminium is. All I know about it are the typical descriptive characteristics that accompany aluminum. When I am introduced to it, suppose that someone says "That's aluminum" and points at it. My experience underdetermines whether or not what I am seeing is aluminum or twaluminum. Every subsequent encounter with aluminum will be similar, because there is nothing that will tell the two metals apart without very close examination (let's say requiring a laboratory). So, how could my mental content about aluminum be different from my counterpart on twin-earth who has just the same sorts of experiences... someone points out to him twaluminum saying "That's twaluminum." He goes on to have the same sorts of experiences I do, but in every case with twaluminum instead of aluminum. What grounds do I have for saying that his concept of twaluminum differs in the least from my concept of aluminum? On the classical sense-datum conception of mental content, there doesn't seem to be any reason to think that there could be any difference between Al's and Twal's concepts, because all of the sensible properties that Al and Twal are aware of are exactly similar phenomenologically. 

The solution that has generally been recommended is externalism about content. That a belief is about a given proposition is not only determined by the thinker, but by the thinker's environment. Putnam seemed to think it was the local linguistic community. It might just as well be determined by causal contact with the things "out there." 

What makes this somewhat disturbing for the traditional epistemologist is something like this: I find the notion that epistemology is a first-person enterprise attractive and think that there are some propositions that I am the expert on, probably to the point of infallibility. If I am in intense pain, there is no way that I can be wrong about thinking that I am. If I am thinking about giraffes, there is some part of that thought that is simply not in doubt. But if content externalism is true, this first-person privelege seems somewhat fantastical. I cannot tell, just by inspecting my own thoughts what I am thinking about. I may not even be able to tell what propositional attitude I am in, if any. For instance, belief requires a proposition, at least on a standard interpretation (we do sometimes talk of believing in a person - but let's set those cases aside). Suppose that propositions have their subjects as constituents, so when I mentally token the proposition expressed by "That canoe is made of aluminum", I express a proposition that has a canoe as a constituent. Suppose when I take myself to token that proposition and believe it, I do it in a case where I do not actually see a canoe, but am the victim of an illusion. Thus, we a case similar to a Frege-puzzle. But unless propositions can have gaps (and they might), then there is properly no proposition that I am believing, so though it seems to me that I am believing a proposition, I am NOT. So, it would seem that I'm not in a believing state at all. But that seems batty in the extreme - how could I be mistaken about that kind of thing? 

[1] Williamson, Timothy, Knowledge and Its Limits...

[2] Burge, Tyler, "Two Thought Experiments Reviewed", Notre Dame Journal of Formal Logic, Vol 23.3 July 1982

[3] Burge, 284.

Stoic Views on Value

In EN 1096a2, Aristotle maintains that no one would call a virtuous person happy if she were suffering great misfortunes, except to defend a “philosopher’s paradox.” The Stoics, however, embrace this paradox by insisting that virtue is sufficient for happiness. But unlike Aristo and his followers, they also hold that so-called external goods, while not choiceworthy in themselves, do have a kind of value in that they are to be preferred to their opposites. However, it remains to be seen whether the Stoics can consistently maintain these two views of value without their position collapsing into either the Aristotelian view or the view of Aristo. It is this tension that Cicero highlights in the dilemma he poses to Cato in De Fin. III.10-12. I will argue that Cato’s subsequent attempt to reconcile these two commitments ultimately fails.

Cato states that what is valuable is that which is “either itself in accordance with nature, or brings about something that is” (III.20) and that the only things worth choosing are things that are valuable. He then describes “appropriate actions” as being impulses directed towards selecting what is in accordance with nature: initially as infants we seek to preserve our own existence, but as we become able to exercise our reason we see an overall order and pattern in the world that we come to seek for its own sake. Thus, we discover that acting virtuously, i.e. acting in accordance with the order and pattern of nature, is the only thing that is truly valuable and thus truly choiceworthy. Virtue is necessary and sufficient for eudaimonia, and all else is indifferent. But of the things that are indifferent, some such as health and beauty are in accordance with nature, and are thus to be preferred over their opposites such as sickness. It seems to me as though Cato is saying that virtue is the only thing we need in order to be happy, but if we are presented with the option of either having or not having certain external goods, we will take those things that are in accordance with nature. So, the Stoic sage will not go out of his way to obtain health and money, but if they happen to come his way he will not reject them. This view of virtue, taken with the Stoic belief that virtue is within our power, shows that for the Stoics happiness is something that is completely within our power, and not dependent on things such as external goods that are often beyond our control.

Cato sees the potential difficulty in making a distinction between what is to be chosen and what is to be selected, and at III.22-24 he tries to illustrate how the Stoics can maintain this distinction while still maintaining that virtue is the only thing that is ultimately valuable. His example of archery is rather confused (why else does the archer practice his craft, if not for the sake of hitting the mark?) but his examples of acting and dancing seem to work better, since they accord somewhat with his account of virtue in that the craft is practiced for its own sake, and not for the sake of some external object. But even dancing and acting are not perfect examples, Cato says, because their success can depend partially on what is outside of the dancer’s or actor’s control, such as the audience’s reaction.

However, it seems to me that the distinction between what is choiceworthy and what is selected cannot be consistently maintained. Earlier in III.20 Cato stated that what is in accordance with nature is valuable, and thus choiceworthy. But at III.44, Cato admits that health has some value, although he does not admit that it is a “good.” He seems to be saying that the value of heath and the value of virtue are different in kind, even if his comparisons of them at III.45 seem to imply a difference in degree. But he is faced with a dilemma: either health is in fact in accordance with nature, and thus is valuable and choiceworthy, or else it is not choiceworthy, and thus can be neither valuable nor in accordance with nature, and thus there would be no grounds for preferring it to sickness. As it is, Cato seems to want to say that health is both in accordance with nature, and yet not choiceworthy, which is inconsistent with his own stated system of value.

A Mereological Argument for Theism

This argument, as far as I know, is novel. It mirrors an argument for the existence of God in Richard Swinburne's The Existence of God, which argues for the existence of God from otherwise inexplicable brain and mental event correlations.

Consider the Special Composition Question: "What necessary and jointly sufficient conditions must any xs satisfy in order for it to be the case that there is an object composed of the xs?"[1]

Ned Markosian has argued that every answer yet suggested for answering this question has failed. He argues that we ought to take composition to be brute: "There is no true, non-trivial, and finitely long answer to SCQ." [2] Let's agree with Markosian that all previous answers to this question have failed and hold off on endorsing his answer to the question.

Suppose however, that none of the views above work, but it is still the case that the some xs are such that the xs compose some y but some zs are such that they do not compose some w. That is to say, some, but not all mereological simples compose something. There should be some fact of the matter about this too. In theory, we should be able to draw up a list of all of the composite objects and their simples with their properties. On brutalism, there is not going to be any explanation for why these simples with these properties compose something, rather than not. The set of correlated xs with their properties and composite ys will be brute, with no explanation.

Brute explanations are only epistemically permissible if there is no uneliminated explanation. [4] I submit that there is an uneliminated explanation. The correlation of xs with their properties and composite ys has a personal cause, a la Swinburne. [5] In fact, whether or not humans are composite physical objects, they will be unable to make it the case that these correlations are true. This is easy to see, there was a time before the first human. It is not plausible to think that the advent of humans started making trees out of xs shaped tree-wise. So, if there is a personal explanation, then it is not a human person. Furthermore, the personal cause must be a mereological simple himself otherwise, he would be insufficient to explain his own composition out of some xs. God fits this description and it is more plausible that the correlations of simples and their properties and the objects they compose are the result of the action of a divine personal cause than that they are brute.

So,

1. There are correlations between simples and composite objects.

2. If there are correlations of this sort, then said correlations are either explained or they are brute.

3. Said correlations are either explained or they are brute. (MP 1, 2)

4. If the correlations are explained, they are either explained by some form of Contact (and all strengthened forms, up to Fusion), Life-ism, Nihilism, Universalism or have a personal explanation.

5. It is false that they are explained by some form of Contact, Life-ism, Nihilism, or Universalism. (Not argued here, between van Inwagen and Markosian every one of these is dispatched). [6]

6. If the correlations are explained, they have a personal explanation.

7. Said correlations either are explained by a personal explanation or they are brute.

8. It is more probable that there is a personal explanation than that they are brute.

9. If there is a personal explanation, then there is a God. (see argument above).

10. It is more probable that there is a God than that the correlations are brute.

Even if this argument works, it has a modest conclusion. As with Swinburne's argument from mental/physical event correlations, it works best as part of a cumulative case.

I'm interested to hear what people think. I have my doubts about premises 5 and 8.

[1] Markosian, Ned, "Brutal Composition" Philosophical Studies 92: 211-249, 1998., 212

[2] Ibid., 214.

[3] Ibid., 223.

[4] I take it that this is plausible. However, I haven't the foggiest idea how to argue for it.

[5] Swinburne, Richard The Existence of God, 2nd Edition(Clarendon Press: Oxford), 2004, 35-51.

[6] van Inwagen, Peter, Material Beings, ( ), 1990.

On the Nature of Numbers

The following definition of the number 2 has been brought to my attention: "[t]he set of all pairs." Against this definition, I raise the following objections:

I. It seems viciously circular: if 2 is the set of all pairs, what are pairs other than a set of two objects? The same objection seems to hold for defining 3 as the set of all triples, etc.

II. Even if this kind of definition can be used for the natural numbers, I am at a loss to understand how numbers such as fractions, irrational roots, pi, imaginary numbers, and negative numbers are to be defined in terms of sets of objects.

III. If “pairs” is understood to mean “pairs of physical objects,” then it would seem that in a possible world with only two physical objects, numbers larger than two could not exist, since e.g. “the set of all triples” would be empty.

IV. Or again, take two possible worlds which each contain only twenty physical objects, with the stipulation that each world contains objects that are entirely different than the objects in the other. In this case, it would seem that “2” in one world would not be identical to “2” in the other, since the set defined as “the set of all pairs” would have different members in each possible world.

Thales' Well

The origin of this blog's name, from Theaetetus:

"[T]hey say Thales was studying the stars... and gazing aloft, when he fell into a well; and a witty and amusing Thracian servant-girl made fun of him because, she said, he was wild to know about what was up in the sky but failed to see what was in front of him and under his feet. The same joke applies to all who spend their lives in philosophy."[1]

This blog will be dedicated to the pursuit of looking aloft and will avoid discussion of what is at our feet.

[1] Plato, Theaetetus, trans. M.J. Levett, rev. Myles F. Burnyeat, from Plato: Complete Works ed. John M. Cooper, Hackett: Indianapolis, IN, 1997., 174a.

Aristotle on Knowledge and Necessity

The determinist’s argument in De Int. 9 (as given by Aristotle and interpreted by Richard Sorabji[1]) runs as follows: it is either true or false that event X will occur. If the statement “X will happen” turns out to be true, then it will have been true in the past. But if the statement “X will happen” was true in the past, then there is no point to deliberating whether or not to bring it about that X, since “X will happen” was true in the past, and the past cannot be altered. Thus, “X will happen” is necessarily true.

As Sorabji points out, Aristotle seems to accept the inference from the premise that any given statement is either true or false to the determinist conclusion. According to what Sorabji calls the traditional interpretation, Aristotle answers the determinist by arguing that predictive statements like “X will happen” are not yet true or false, but become true or false when X actually happens (or becomes inevitable). Sorabji finds this response unsatisfying, because e.g. if a statement’s being true means that it corresponds to a state of affairs in the world (which was a fairly common understanding of “true” for the Greeks) then the statement “there will be a sea battle tomorrow” is true (when uttered) if in fact it corresponds to a state of affairs, viz. the sea battle happening tomorrow; to speak of it as “not yet true” seems odd (Sorabji 101).

Sorabji’s own solution is to deny the inference between the premise that any given statement is either true or false to the determinist conclusion by making reference to one’s power, or lack of power, to bring about a state of affairs. Thus, if the statement “X will happen” turns out to be true, i.e. actually happens, then I do not have it in my power to make it the case that not-X, since X has already happened and I cannot alter the past. But, Sorabji thinks that this scenario is disanalogous to a situation in which I have been told “X will happen,” but X has not yet happened. Sorabji argues that since the event X still lies in the future, we still have the power to bring about not-X, and thus even if the statement “X will happen” does turn out to be true, it is not necessarily true, as the determinist thinks.

However, it seems to me as though something like Sorabji’s reply was addressed by Aristotle. At 19a7-23, Aristotle argues that it is false that all events happen of necessity: before an event X happens, both X and not-X are possible states of affairs. Since not-X is possible, then if X happens, it cannot have happened necessarily since it was in some agent’s power to bring about not-X. This passage seems to mirror Sorabji’s discussion of irrevocability in the first paragraph on p. 102, and Aristotle seems to argue here for the same disanalogy between past events and future events that Sorabji does.

Returning to Sorabji’s objection that it seems as though statements about the future should be said to be actually true or false depending on whether they in fact correspond to a state of affairs, it seems to me that Aristotle could answer that they are not yet true or false simply because they do not yet correspond to a state of affairs, and that only when they can truly be said to correspond, or fail to correspond, to a state of affairs can they be said to be true or false. It seems to me that this is the line Aristotle takes in his conclusion to this section at 19b1. Thus it appears that Aristotle argues both for a qualified principle of bivalence (cf. Sorabji p. 94-95) and against the determinist’s belief that all events happen of necessity.

---

[1] Sorabji, Richard. “Necessity, Cause, and Blame: Perspectives on Aristotle’s Theory.” Duckworth, 1980, pp. 91-103.

A Phenomenological Challenge for Evidentialism

"I think that the so-called analogical argument for the existence of other minds is a fairly good argument - as philosophical arguments go - but my belief that my wife has thoughts and sensations generally similar to my own is not based on this or any other argument. Not even partly."[1]

A friend of mine turned me onto this paper by PVI. It's very interesting, and fits in with some reading I'm doing for a seminar on rational disagreement. It's a fascinating paper - I enjoyed reading it - but I almost always enjoy reading PVI, whether or not you agree with him on a given point, he's a delightful writer.

At any rate, I see a challenge for evidentialism here. I mean by 'evidentialism' something like the following: p is justified for S at t iff p fits S's evidence E at t (cribbed directly from Conee and Feldman's Evidentialism). In the context, what PVI is talking about beliefs that we are "hardwired" for, beliefs that we come to "automatically" without having to resort to discursive reasoning.

I think there is also something to what PVI is saying - phenomenologically, there are many beliefs we arrive at that we don't obviously carefully survey our evidence, weigh it, and believe. Belief is frequently automatic. My belief in other minds is one of these things - if I walk into the grad office and see two students sitting there, I don't make a reasoned argument that I am aware of to the conclusion that by certain analogies with my own mind and behavior, the two objects in the room are actually intelligent human beings with minds more or less like my own. I think that PVI might be right also about the material world. I don't wake up in the morning, open my eyes, look out the window and conclude that the simplest explanation is that the things that I am experiencing are non-mental objects independent of my mind. However, I believe it.

So, there's a phenomenological challenge to evidentialism. Evidentialists need a way to account for automatic beliefs that seem to be justified. But I think that the problem is hardly one for evidentialism simpliciter, as an account of justification, but rather with evidentialism as part of the analysis of knowledge. The way I'm inclined to go with this (at least at the moment) is to take the automatic part of belief formation to be part of a basing relation that either causally or otherwise pairs beliefs and evidence (cribbed again from Conee and Feldman).

Aside: for an access internalist, is it necessary that both the justifier of a belief and the basing relation be accessible to the knower?

[1] van Inwagen, Peter, "Is God an Unnecessary Hypothesis?" God and the Ethics of Belief ed. Andrew Dole and Andrew Chignell (Cambridge) 2005, p. 146