Inverting the stack

This is probably hopelessly outdated, but back when I was more up to date with the hot new programming paradigms and the like, people liked talking about “web stacks,” with the archetypal one being LAMP. At the bottom you had the operating system (Linux), then the server software (Apache) on top of that, then the database software (MySQL) on top of that, and then the programming language that you write the site’s code in (Perl/Python/PHP, depending on your preferences) operating over everything.

The vertical metaphor is useful for pretty obvious reasons. You can’t get much done writing a website with a given language without data in a database for your code to play with. There’s no point having the data on a server without server software with which to make it accessible to users. And you can’t run anything without your OS underneath it. Each layer needs the layers underneath it to do any work.

I wonder if it’s useful to think about philosophical frameworks, or at least those which Rorty would have derisively categorized as “systematic philosophy,” in a similar fashion. At the bottom, you have metaphysics to tell you what stuff exists. Then you have epistemology on top of that to tell you what you can know about that stuff. Then you have philosophy of science (and of the various sciences, and perhaps philosophy of mind and language too) to tell how you can know the stuff that your epistemology says you can know. And finally you have practical philosophy — ethics, political philosophy, aesthetics, and so forth — to tell you what you should do in the world that the rest of philosophy has laid out for you.

Very few people this side of Kant lay out a full stack, but Cornell Realists like Richard Boyd and Nicholas Sturgeon come pretty close. At the very bottom you have metaphysical naturalism: what exists is the natural, physical world as described by modern science. Epistemologically you have some kind of reliabilism: you are justified in believing what reliable processes (to whit, the scientific method) lead you to believe. Methodologically, you have scientific realism: science is the best means of acquiring knowledge. And then you have naturalistic utilitarianism on top of it: you should do what maximizes the good, and “good” is just the natural fact of maximizing utility. It looks sort of like this:


I don’t presume to speak for anyone but me, but this is certainly how have usually thought about philosophy. And I’m starting to think, for reasons I touched on two years ago here and in part as a result of mulling of Matt’s response, it’s kind of mixed up. There is no necessary reason for the stacking to go in that order. Indeed, it might make more sense to make value theory foundational.

Consider this stack:


Instead of starting from the foundation of naturalism, someone who accepts this stack starts with his ethical theory — in this case utilitarianism, for ease of illustration — and moves from there. The question “what beliefs are justified?” is rephrased as “what ought I believe?” and so out comes a kind of pragmatism (“one ought to believe whatever beliefs maximize utility”) as one’s epistemological theory. And out of that comes an instrumentalism about science; it’s a tool for figuring out what’s best to believe, not a means of access to a mind-independent reality. And as for metaphysics — whatever! You don’t really need it. You have a pretty coherent picture of what one should do and believe and think about the world without getting into ontological matters.

The obvious objection here is, “But how do you know utilitarianism is true?” From the perspective laid about above, however, this reply is sort of confused. Utilitarianism isn’t something you can know. It’s a precondition of obtaining knowledge. If to “know” something consists in believing what you ought to believe about a given subject, then you can’t explain how you know something without making reference to an underlying ethical (or at least normative) assumption about what’s best to believe.

You can think about this as a kind of fideism about ethics. Just as theological fideists don’t think that it’s necessary or even wise to come up with reasons for believing in God, on this view you can’t come up with reasons to believe one ethical doctrine or another. All you can do is take your stand and see what follows.

I don’t necessarily buy this view, though as you can probably tell I’m quite sympathetic to it. And one reason for that is that it avoids a difficulty that my previous post was trying to identify in the Cornell version of utilitarianism. On Boyd’s view, the answer to “should I perform action A?” is always “you should if A maximizes utility” except if action A is of the form “hold belief X”. Answers to questions like “should I hold belief X?” are more like “if belief X is verified by scientific methods and corresponds to the mind-independent world.” But why should this kind of question and only this kind of question be exempt from utilitarian reasoning? Isn’t it simpler to just junk the separate epistemological theory and start with the ethical theory as a foundation?

Steve Petersen makes a persuasive case along these lines from an act utilitarian perspective here. As Petersen says, the burden of proof should be on value pluralists to explain why epistemic value is special and deserving of its own treatment apart from all other kinds of value. This challenge is particularly grave if, like Cornell realists, you think that epistemic value is the only exception to your view that one kind of value holds for all things.

P.S. Paul Kelleher, who unlike me is an actual philosopher (with a PhD from Cornell, no less!), notes that it’s probably fairer to characterize the Cornell folks as non-utilitarian consequentialists. D’oh. In any case, I think the above critique still applies.

Joyce on Skepticism

I must confess that I grow weary of attacking moral naturalism. Speaking as both a moral skeptic and an atheist, I find myself classifying defenders of moral realism along with apologists for theism (and I have never bothered to argue against theists). Both, to my mind, have about them an air of slightly desperate conservatism: an anxious determination to ensure that popular belief systems turn out as true. I do not accept as a general rule the orthodox methodological principles underlying such an approach; I do not think it the job of the philosopher to leave ordinary beliefs and attitudes as unruffled as he or she can. How much more invigorating philosophy might be if it ruffled us; how much more intriguing life might be if we opened our minds to the possibility that we’ve all been dramatically mistaken about the nature of the world.

Rest here. See also Guy Kahane’s response here.

Neil Sinhababu on “Unequal Vividness and Double Effect”

Here’s the summary of the argument (links added by me for context):

I hope to advance this debate with an argument that uses recent empirical results against the Doctrine of Double Effect as Greene and Singer wish to, but which avoids the problems Berker notes, and which better fits the data. It runs as follows:

[Unequal Vividness Explanation] Double Effect is accepted because of how unequally vivid representations of actions’ intended and merely foreseen consequences affect our desires.

[Unreliability Claim] When the effects of unequally vivid representations upon our desires are decisive in causing our judgments about what to do, we are usually mistaken.

[Conclusion] In accepting Double Effect, we are likely to be mistaken.

Full paper here. What’s cool about this is that the unreliability claim, as I understand it, rests on both a descriptive claim (in practical judgment, unequal vividness frequently leads to error) and a normative claim (error in practical judgment is likely to constitute error in moral judgment). So I don’t think Neil is trying to derive an is from an ought. But at the same time, the normative claim in question is extremely modest, and one that most people involved in the dispute in question should be able to accept. The method he’s using, then, seems capable of using neuroscience to make progress in previously intractable disputes like this without falling victim, as Greene does, to errors that Hume was pointing out three centuries ago.

Unequal vividness also seems like a promising way of deflating other deontological hobby-horses, like the causing/allowing harm distinction. Surely one reason that we’re less willing to condemn someone for failing to donate $10 to UNICEF, even with the knowledge that the donation will likely save a child’s life, than we are to condemn someone who runs over a child with his car and then keeps going is that the latter scenario is far more vivid.

Anti-Gridlock Amendment

Reading through Slate‘s forum on changing the Constitution, I was rather struck by the absence of proposals intending to move toward a system more typical of developed countries. Sure, I wasn’t expecting full-on parliamentarism (the Senate and the presidency are abolished, executive functions are performed by the Speaker of the House, etc.). But I was at least expecting the issue to come up.

There are the uncontroversial measures: moving up inauguration, establishing a formal right to vote, etc. There are the good government measures that are nonetheless pretty small bore, like Supreme Court term limits, electing the attorney general, and depoliticizing election management. There are, of course, the proposals to reverse Citizens United. There are the policy proposals that sounds more like bills than amendments. Mike McConnell’s proposal for changing the procedure for setting Congressional rules, so as to make filibuster reform easier, is the only one that even tries to make it easier to pass legislation, but it would make things more like pre-Bush/Obama America than Western Europe.

Some of this, I’m sure, is due to professional self-selection. If you spend your life studying the Constitution, you probably (unless you’re Sandy Levinson) have a certain affection for the document and its idiosyncrasies, including those that make governing in the presence of a strong party system exceedingly difficult. The fact that the most dramatic proposals along these lines, like Senate abolition or letting the president dissolve Congress and call new elections, require Congress to vote to reduce its own powers probably doesn’t help their chances of inclusion either. But the difficulty of forming a government seems rather obviously like the biggest Constitution-level problem of the moment, and so it’s strange to see it given short-shrift. When passing a budget annually is no longer politically possible, and Congress’ disagreements with the president almost lead to default on a regular basis, something has gone badly awry.

In any case, here’s a relatively modest proposal along these lines, inspired by proposal one from Lloyd Cutler’s “To form a government”, which I think is the likeliest to pass of anything that would seriously help:

Section 1. The House of Representatives shall be composed of Members chosen every sixth year by the People of the several States.

Section 2. The President and Vice President shall be elected jointly by the direct vote of the citizens of the United States every sixth year, without regard to whether the citizens are residents of a State.

Section 3. The election of all Members of the House of Representatives and of all Senators shall coincide with the election of the president.

Section 4. The Congress shall have the power to enforce this article by appropriate legislation.

States could then decide whether the two Senators are to be elected simultaneously in separate races (like when Chuck Schumer and Kristen Gillibrand were both up for reelection in 2010), or through single transferable vote or some other multiple-winner method.

There are a few advantages to this approach. House members would be inclined to support it to get longer terms, and Senators wouldn’t lose anything. While synchronizing elections would almost always mean the same party would control both chambers and the presidency, voters would still have the option of choosing divided government, so checks and balances junkies wouldn’t be able to complain. It doesn’t solve the geographical bias of the Senate, and a unicameral parliament would be simpler, but it’d all but eliminate gridlock without threatening the members of Congress who’d have to vote for it.