In 1962, a short philosophy paper caused a little flurry in philosophical circles. Two decades later David Foster Wallace, armed with further developments in analysis, created an elaborate system of notation to solve the problem raised in the paper. This book contains Richard Taylor’s original paper, the resultant flurry, and Wallace’s solution. It also contains a fair amount of background information about the whole exchange.
I think most David Foster Wallace fans who pick up this book would enjoy the background info, and may be interested in the philosophical back-and-forth sparked by Taylor’s paper. Probably the least engaging part is actually the bit written by Wallace. It’s not surprising that he speaks the language of academic philosophy like a native; but that can be a hard language for the outsider to penetrate at times. This book is subtitled “An Essay on Free Will” but it really is more a technical paper (it’s a thesis he wrote while at university). The title of section V, A formal device for representing and explaining the Taylor inequivalence: features and implications of the intensional-physical-modality system J, gives a good flavour of the piece’s uncompromising nature. And you thought Infinite Jest was dense.
So the whole thing started back in the ’60s when the philosopher Richard Taylor came up with a ridiculous and somewhat convoluted argument purporting to show that, given certain reasonable-sounding assumptions, fatalism was true; in other words, we do not have control over our destiny. A simpler version of the argument could run as follows:
- Tomorrow, either I will wear my hat or I won’t wear my hat.
- If I will wear my hat tomorrow, then I can’t refrain from wearing my hat tomorrow; so I don’t have a choice whether to wear my hat.
- If I won’t wear my hat tomorrow, then I can’t wear my hat tomorrow; so I don’t have a choice whether to wear my hat.
- Both these cases lead to the same conclusion.
- Therefore I don’t have a choice whether to wear my hat.
The same argument could be made for any other decision I could make. So therefore, I don’t have any choice at all about anything. Fatalism is true. OMG FML.
Is this valid? Yes, I would say so. But is it fishy? Yes, definitely. And the reason is that it is not sound: the premise on the first line is false.
As interpreted on line 2, to say that “tomorrow I will wear my hat” is already to assume that it is certainly going to happen, that is to say neither I nor anyone else has any choice about the matter. And the same is true about “tomorrow I won’t wear my hat”. That’s why in each case we can conclude that we have no choice in what happens. It’s an unusually (perhaps unrealistically) strict interpretation of the word “will”. Under this interpretation, yes, line 1 implies fatalism. But that doesn’t matter because line 1 is not true.
We can accept as an axiom that every statement is either true or false. Therefore, either tomorrow I will wear my hat, or it is false that tomorrow I will wear my hat. But this is not the same as line 1. To show that it is, we would have to show that “it is false that tomorrow I will wear my hat” implies “tomorrow I won’t wear my hat”. This implication is assumed by Taylor, and by all his critics, and causes everybody to tie themselves up in knots looking for problems elsewhere in his argument. But it’s just a very basic problem in modal logic: false(tomorrow(wear my hat)) does not necessarily imply tomorrow(false(wear my hat)). To show that it did, you would need to come up with some startlingly unorthodox arguments; Taylor doesn’t even try. To be fair, I don’t think modal logic was really a thing until the 1970s, but it’s not that complex (say I, 50 years later) and Taylor was a professional philosopher.
Actually, I have just realised that my argument above is essentially the same as Wallace’s. His preliminary statement in section III on page 164 is the same as the above, adapted to the (needlessly) more complex argument offered by Taylor. Wallace dots every i and crosses every t in his paper, compared to my hand-waving approach, which explains why his thesis is 80 pages long rather than one paragraph.
Anyway, once the above is established, we can see that any contingent statement about the future must be false. It is false that I will wear a hat tomorrow, and it is also false that I won’t wear a hat tomorrow. This seems counterintuitive, but that’s because normal usage of the future tense is a bit woolly (like my hat); we have to tighten up our meanings in order to draw rigorous conclusions. So:
- “Tomorrow, I will wear a hat” is false. (It can’t be true (yet) because it’s not tomorrow (yet).)
- “Tomorrow, I won’t wear a hat” is also false. (It can’t be true (yet) because it’s not tomorrow (yet).)
- Therefore “Either tomorrow I will wear a hat or tomorrow I won’t wear a hat” is false
- In other words “Tomorrow, either I will wear my hat or I won’t wear my hat” (that’s line 1) is false
- Therefore line 1 is false.
- Therefore Taylor’s argument falls ignominiously at the first hurdle.
One of the reasons none of Taylor’s critics take this position is that Taylor’s original argument is more long-winded, and drags in concepts of causality and counterfactuals. This all serves to obscure the basic problem with “will” that slips through and muddies the waters.
Wallace spots the fundamental error in Taylor’s argument, but I feel he overcooks his analysis of it. It’s as if he has noticed his house is cold, so installs an electronically-controlled central heating system to fix the problem; not noticing that the front door was wide open all the time. He’s just being extremely thorough, though he could have simplified things if he’d started off with a slimmed-down version of Taylor’s argument like mine. The intervening 20 years of philosophical development probably helped, but even so he’s certainly more thorough than all the other philosophers who responded to Taylor’s original paper.