Now there are one hundred things to be classified: eighty topics and twenty subtopics. The subtopics are individually easier to classify, so that’s nice. On the other hand, I started with a hard task of classifying ninety things, and now I have the hard task of classifying one hundred things, and this doesn’t look like progress.
So it’s time to introduce the tool that I primarily used as a guide to classification. It’s another kind of binary sort, like I used in the last section. Except now instead of applying it to all the articles in one topic, I apply it to all the articles in two categories. The intuition here is that if I’ve really got a good categorization, all the topics/subtopics within each of the categories should cluster together. If the topics/subtopics are not clustering, then that is a reason to be sceptical of the categorization.
Here is a clean example of how it might work. I generated a two topic LDA out of all the articles in the categories ethics and metaphysics. And for each topic, I asked what the average probability was that the article was in topic 1. Remember which topic gets labelled 1 is arbitrary, so this is just asking how close, on average, each of them was to one arbitrarily chosen end of the binary sort.
|Ethics||Promises and imperatives||0.138|
|Ethics||Abortion and self-defence||0.175|
|Metaphysics||Classical space and time||0.962|
|Metaphysics||Composition and constitution||0.970|
That works pretty well. There is a very clean split between the ethics subjects and the metaphysics subjects. (I’ll use “subject” from now on to refer to both the topics and the subtopics that are being classified.) The only one vaguely in between is Frankfurt cases, and that even makes sense; until recently free will was a subject in metaphysics textbooks as much as ethics textbooks.
Five caveats before I continue.
First, which category bunches near 0 and which bunches near 1 is completely arbitrary and doesn’t mean anything. What matters is the bunching.
Second, the bunching matters much more than which is on either side of 0.5. Hopefully there will be a big gap in the means somewhere, and that will correspond to a category boundary. But especially when the categories are of different sizes, that gap might be a long way from the midpoint.
Third, this is just a tool. There are going to be cases it gets wrong, and I’ll correct them by hand. But it’s a surprisingly powerful tool, and I’ll defer to it in a lot of close calls.
Fourth, I didn’t just find these categories lying around and used this tool to confirm them. There was a lot of juggling around to get it to a point where most of these automatic classifications agreed with my classifications, and the ones that didn’t were easy enough to explain. And the nature of LDAs is that any change somewhere creates changes everywhere. The methodology here involves a fair bit of trial and error. But some of the methodology is clear enough. Look back at that table. If I’d put Frankfurt cases in metaphysics, the very same test would tell me to move them into ethics. So this method doesn’t just confirm that a classification gets things broadly right, it can say where a classification is going wrong. To be sure, it’s basically an equilibrium method, and it doesn’t rule out other equilibria. But it’s interesting to have found even one.
Fifth, I’m not applying this to three categories: idealism, aesthetics and philosophy of religion. Both aesthetics and philosophy of religion really don’t have borderline cases. And the technique I’m using doesn’t work so well when the categories are of very uneven size. Therefore, it often gives wonky results for those two categories. (It also on occasion gives some of the cleanest splits - but it feels random when it does.) Still, I wasn’t worried about those two categories. And idealism is a special case that I’ll come back to at the end of the chapter.
I’m looking at nine categories to check that the boundaries between them are drawn roughly correctly. And that means there are nine choose two, i.e., thirty-six, boundaries to look at. These are in subsections, not sections, so they don’t show up in the sidebar. I’ve included links here to the list of boundaries that I’ll survey.
- Epistemology vs. Ethics
- Epistemology vs. History of philosophy
- Epistemology vs. Logic and mathematics
- Epistemology vs. Metaphysics
- Epistemology vs. Philosophy of language
- Epistemology vs. Philosophy of mind
- Epistemology vs. Philosophy of science
- Epistemology vs. Social and political
- Ethics vs. History of philosophy
- Ethics vs. Logic and mathematics
- Ethics vs. Metaphysics
- Ethics vs. Philosophy of language
- Ethics vs. Philosophy of mind
- Ethics vs. Philosophy of science
- Ethics vs. Social and political
- History of philosophy vs. Logic and mathematics
- History of philosophy vs. Metaphysics
- History of philosophy vs. Philosophy of language
- History of philosophy vs. Philosophy of mind
- History of philosophy vs. Philosophy of science
- History of philosophy vs. Social and political
- Logic and mathematics vs. Metaphysics
- Logic and mathematics vs. Philosophy of language
- Logic and mathematics vs. Philosophy of mind
- Logic and mathematics vs. Philosophy of science
- Logic and mathematics vs. Social and political
- Metaphysics vs. Philosophy of language
- Metaphysics vs. Philosophy of mind
- Metaphysics vs. Philosophy of science
- Metaphysics vs. Social and political
- Philosophy of language vs. Philosophy of mind
- Philosophy of language vs. Philosophy of science
- Philosophy of language vs. Social and political
- Philosophy of mind vs. Philosophy of science
- Philosophy of mind vs. Social and political
- Philosophy of science vs. Social and political
|Ethics||Abortion and self-defence||0.348|
|Ethics||Promises and imperatives||0.794|
I’m doing these in alphabetical order, and that means the first cab off the rank is one of the trickiest. Three of the epistemology topics: formal epistemology, knowledge and justification are clear enough. And most of the ethics topics are clear enough. But there are two that are hard.
One is arguments. Why should this go with epistemology? Everyone uses arguments. I’ve put it in epistemology for three reasons. First, as you can see, the model puts it there. (And we’ll keep seeing that as we look through the comparisons between epistemology and other categories.) Second, once you extract the dualism arguments from the topic, what’s left are primarily papers about what we can learn from arguments. And those feel like epistemology papers to me. And third, there are a few other close calls where I put something that could go in epistemology elsewhere, and getting the overall shape of the graphs right felt like I needed to have one close call go this way.
The other strange one is decision theory. It’s easy to think that would simply go with the other probability papers in epistemology or maybe philosophy of science. But it’s ended up in ethics for two reasons. One is that, as the table shows, that’s where the automatic sorter put it. But the other comes from thinking about what articles are left in that topic. Given that formal epistemology exists as a topic, and that theory testing also exists, there isn’t as much directly about probability in this topic. What is left is primarily papers about value functions. They are very technical questions about value functions, to be sure, but the papers actually in this topic on the whole are more about the “value” part of expected value than the “expected” part. And that isn’t absurd to group in with ethics. Put another way, what we really have in this topic is formal ethics, and it makes sense that goes with ethics.
|History of philosophy||Hume||0.644|
|History of philosophy||Dewey and pragmatism||0.806|
|History of philosophy||Ancient||0.842|
|History of philosophy||Early modern||0.902|
|History of philosophy||Social contract theory||0.912|
|History of philosophy||Kant||0.916|
|History of philosophy||Other history||0.948|
|History of philosophy||Heidegger and husserl||0.971|
History is a fairly heterogenous category, and this technique doesn’t work as well with it as the other categories. But it is fairly happy with where the episteology/history boundary is drawn.
|Logic and mathematics||Vagueness||0.691|
|Logic and mathematics||Verification||0.741|
|Logic and mathematics||Deduction||0.798|
|Logic and mathematics||Truth||0.813|
|Logic and mathematics||Analytic/synthetic||0.841|
|Logic and mathematics||Propositions and implications||0.854|
|Logic and mathematics||Definitions||0.936|
|Logic and mathematics||Mathematics||0.950|
|Logic and mathematics||Universals and particulars||0.964|
|Logic and mathematics||Sets||0.964|
I would have expected the model would have really wanted to put arguments in with logic and mathematics. And it is a bit of a borderline case. But it still clearly puts arguments closer to the epistemology cluster than the logic cluster. And I was surprised that verification didn’t look a bit more like an epistemology topic. I’ll come back to why it ends up in logic and mathematics not anywhere else, but for now it looks like it shouldn’t be an epistemology topic.
|Metaphysics||Composition and constitution||0.038|
|Metaphysics||Classical space and time||0.047|
A lot of philosophers use the phrase “metaphysics-and-epistemology” almost as if it is one long word, with “em-and-ee” being shorthand for an alleged field within philosophy. We don’t really see any such field turning up in this model. The gap between metaphysics and epistemology is as clear as any gap between categories.
|Philosophy of language||Language norms||0.364|
|Philosophy of language||Belief ascriptions||0.678|
|Philosophy of language||Speech acts||0.787|
|Philosophy of language||Radical translation||0.846|
|Philosophy of language||Meaning and use||0.924|
|Philosophy of language||Sense and reference||0.936|
|Philosophy of language||Denoting||0.963|
Here is the first occasion I’ve had to overrule the model. It wants to put language norms in with epistemology not philosophy of language. And it could go either way. But I’ve put it with philosophy of language for a couple of reasons. One is that the paradigm articles in this subtopic, which are often by or about Brandom, feel more like language articles than epistemology articles to me. And the other is that when I ran the models with this shifted to epistemology, a lot of the neat divisions we’ve already seen got less neat. Still, this is one of the trickier classifications, and I suspect with another few years of data, the model would have found a norms topic that was properly divided into ethics and epistemology.
|Philosophy of mind||Conceivability arguments||0.187|
|Philosophy of mind||Intention||0.464|
|Philosophy of mind||Freud||0.500|
|Philosophy of mind||Olp mind||0.503|
|Philosophy of mind||Concepts||0.697|
|Philosophy of mind||Wide content||0.742|
|Philosophy of mind||Emotions||0.753|
|Philosophy of mind||Self-consciousness||0.794|
|Philosophy of mind||Minds and machines||0.840|
|Philosophy of mind||Perception||0.893|
|Philosophy of mind||Cognitive science||0.901|
|Philosophy of mind||Color/colour||0.903|
|Philosophy of mind||Physicalism||0.916|
|Philosophy of mind||Psychology||0.978|
There are two tricky cases here, and the model has relatively firm opinions on one of them.
One is what to say about conceivability arguments. The binary sort I just ran is not completely sure, but it prefers to put it back in with epistemology. I didn’t do that for a few reasons. One is that the subject matter sure looks like philosophy of mind to me. If papers on zombies aren’t fin de siècle philosophy of mind, I don’t know what is. The other is that there are several reasons to think the model might have gotten confused here. It isn’t surprsing that a technique that relies entirely on string matching puts the knowledge argument in with epistemology. And that’s doubly so when we conceive of epistemology as including the general study of arguments. So this looks like a philosophy of mind topic, and there are reasons to think the model won’t be smart enough to see this. Therefore, I put it in philosophy of mind.
The other tricky case is perception. Or at least I thought it was tricky. When I was trying to sort the topics manually, I had no idea what to do with it. But the model doesn’t have any doubts at all, and I was happy to let it resolve my uncertainty.
|Philosophy of science||Space and time||0.035|
|Philosophy of science||Quantum physics||0.054|
|Philosophy of science||Evolutionary biology||0.067|
|Philosophy of science||Dna||0.069|
|Philosophy of science||Thermodynamics||0.074|
|Philosophy of science||Chemistry||0.084|
|Philosophy of science||Models||0.112|
|Philosophy of science||Functions||0.218|
|Philosophy of science||Mechanisms||0.277|
|Philosophy of science||Explanation||0.368|
|Philosophy of science||Grue||0.443|
|Philosophy of science||Theories and realism||0.475|
|Philosophy of science||Laws||0.536|
|Philosophy of science||Teleology||0.640|
|Philosophy of science||Methodology of science||0.682|
|Philosophy of science||Chance||0.728|
|Philosophy of science||Theory testing||0.734|
|Philosophy of science||Game theory||0.758|
As someone whose earliest philosophical work sat right on the boundary between epistemology and philosophy of science, I thought that there would be more borderline cases here. But the model wasn’t budging. It’s really certain it wants to put the four epistemology subjcts at one end. And the closest subject to them, game theory, is one that I would have overruled and put back in philosophy of science if it hadn’t got it right. So this looks like a clear split.
I’ve mentioned this before, but I keep being surprised at how well the model separated theory testing from formal epistemology. These have a lot of overlap, and the model does think theory testing is closer to epistemology than most things in philosophy of science. But ultimately it knows how to sort articles into one or the other. I find that remarkable.
|Social and political||Liberal democracy||0.039|
|Social and political||Egalitarianism||0.046|
|Social and political||Marx||0.048|
|Social and political||Feminism||0.048|
|Social and political||Life and value||0.055|
|Social and political||History and culture||0.073|
|Social and political||Political freedom||0.132|
|Social and political||Law||0.198|
|Social and political||Race||0.215|
|Social and political||War||0.294|
Maybe if we ran the tape forward to 2020 there would be more papers on the boundary between epistemology and social and political. Social epistemology has, after all, become a thing. But restricting attention to these twelve journals, up to 2013, this frontier wasn’t heavily populated.
|Ethics||Abortion and self-defence||0.105|
|Ethics||Promises and imperatives||0.151|
|History of philosophy||Social contract theory||0.435|
|History of philosophy||Hume||0.763|
|History of philosophy||Kant||0.813|
|History of philosophy||Dewey and pragmatism||0.829|
|History of philosophy||Ancient||0.922|
|History of philosophy||Other history||0.939|
|History of philosophy||Heidegger and husserl||0.964|
|History of philosophy||Early modern||0.982|
A little bit of overlap here, because social contract theory feels as much like a ethics topic to the model as a History topic. But we know that it goes best with the other History topics. I’m not sure why value feels so historical to the model though; it could be that it has so many older papers in it, but that’s not a great reason.
|Logic and mathematics||Mathematics||0.015|
|Logic and mathematics||Sets||0.020|
|Logic and mathematics||Truth||0.031|
|Logic and mathematics||Universals and particulars||0.038|
|Logic and mathematics||Propositions and implications||0.041|
|Logic and mathematics||Analytic/synthetic||0.052|
|Logic and mathematics||Deduction||0.057|
|Logic and mathematics||Vagueness||0.062|
|Logic and mathematics||Definitions||0.067|
|Logic and mathematics||Verification||0.097|
|Ethics||Promises and imperatives||0.799|
|Ethics||Abortion and self-defence||0.917|
This one is pretty easy. Note that although the model is a little uncertain about decision theory, it is happy to include it in ethics. This is notable because so much of decision theory is about problems about infinity, so it would not have been a surprise for it to go with logic and mathematics.
|Ethics||Promises and imperatives||0.138|
|Ethics||Abortion and self-defence||0.175|
|Metaphysics||Classical space and time||0.962|
|Metaphysics||Composition and constitution||0.970|
This is the one I used as an example, and it seems there aren’t really any borderline cases.
|Philosophy of language||Denoting||0.006|
|Philosophy of language||Sense and reference||0.019|
|Philosophy of language||Radical translation||0.025|
|Philosophy of language||Belief ascriptions||0.049|
|Philosophy of language||Meaning and use||0.056|
|Philosophy of language||Speech acts||0.090|
|Philosophy of language||Language norms||0.268|
|Ethics||Promises and imperatives||0.754|
|Ethics||Abortion and self-defence||0.913|
Nothing too surprising here. The contemporary work on slurs might eventually generate a topic that’s a borderline case. Before I thought of dividing topics in two I was somewhat tempted to classify ordinary language philosophy as philosophy of language. We can see here how bad a mistake that would have been.
|Philosophy of mind||Perception||0.015|
|Philosophy of mind||Color/colour||0.018|
|Philosophy of mind||Physicalism||0.021|
|Philosophy of mind||Wide content||0.025|
|Philosophy of mind||Psychology||0.043|
|Philosophy of mind||Cognitive science||0.052|
|Philosophy of mind||Minds and machines||0.081|
|Philosophy of mind||Concepts||0.092|
|Philosophy of mind||Olp mind||0.130|
|Philosophy of mind||Self-consciousness||0.178|
|Philosophy of mind||Emotions||0.436|
|Philosophy of mind||Intention||0.471|
|Philosophy of mind||Freud||0.537|
|Philosophy of mind||Conceivability arguments||0.558|
|Ethics||Promises and imperatives||0.876|
|Ethics||Abortion and self-defence||0.882|
Another clean division, though the gap between the mind topics and the ethics topics was surprisingly small.
I thought intention, at least in the form that it is in these journals, was clearly a philosophy of mind topic. It’s about a special kind of mental state. But it is fairly common to classify it in with ethics. And it is so clearly tied to action theory (which is more or less in ethics), and the doctrine of double effect (which is clearly in ethics) that you can see why. But I still want to treat papers about a distinctive mental state as being in philosophy of mind, and the model more or less agrees with me.
|Ethics||Abortion and self-defence||0.061|
|Ethics||Promises and imperatives||0.062|
|Philosophy of science||Game theory||0.345|
|Philosophy of science||Teleology||0.477|
|Philosophy of science||Mechanisms||0.768|
|Philosophy of science||Methodology of science||0.782|
|Philosophy of science||Laws||0.809|
|Philosophy of science||Functions||0.839|
|Philosophy of science||Explanation||0.884|
|Philosophy of science||Chance||0.919|
|Philosophy of science||Theory testing||0.924|
|Philosophy of science||Theories and realism||0.943|
|Philosophy of science||Evolutionary biology||0.947|
|Philosophy of science||Chemistry||0.960|
|Philosophy of science||Grue||0.963|
|Philosophy of science||Thermodynamics||0.968|
|Philosophy of science||Models||0.969|
|Philosophy of science||Dna||0.975|
|Philosophy of science||Space and time||0.987|
|Philosophy of science||Quantum physics||0.991|
I didn’t think of the ethics/philosophy of science boundary as being a particularly hard one to identify. But this ended up being trickier than I expected.
One issue is that teleology has enough language that goes with action theory that it ends up confusing the model.
But the other thing is what to do with decision theory and game theory. The model clearly doesn’t like breaking them up. But the material on game theory that is here is primarily evolutionary game theory as used in philosophy of biology; that’s clearly philosophy of science. And it would be a stretch to say that decision theory, which is well under 0.5, is a philosophy of science topic on this basis. So I think the overall best thing to do is what I actually did do—though I can see why others might prefer something else.
|Ethics||Promises and imperatives||0.048|
|Ethics||Abortion and self-defence||0.169|
|Social and political||Political freedom||0.533|
|Social and political||War||0.614|
|Social and political||Race||0.691|
|Social and political||Law||0.753|
|Social and political||Egalitarianism||0.754|
|Social and political||Feminism||0.756|
|Social and political||Liberal democracy||0.883|
|Social and political||Life and value||0.887|
|Social and political||Marx||0.900|
|Social and political||History and culture||0.929|
Here is where I really appreciated having a model to work with. I still didn’t entirely go along with what the model suggested, but it helped see what were the easy cases and what were the hard cases.
If I had to guess I would have put abortion and self-defence in with ethics and feminism in with social and political. But I’m very glad to have a model, not just a guess, to rely on here. And I would have made the same division with the two topics closely connected to Parfit’s work: population ethics and egalitarianism. But again, I would have been nervous about relying on guesswork, and it was nice to see the model agree.
So what to say about the two subtopics in the middle: political freedom and forgiveness? I think these are in practice reasonably clear cases. If political freedom isn’t a topic in social and political philosophy I don’t know what is. And while there are some papers about social and structural matters in forgiveness, it is enough about individual relations that I think it should go in ethics.
I suspect what’s happened with both of these is that the subtopics aren’t as cleanly separated as they appear. There are enough papers about law in the forgiveness subtopic that the model won’t quite push it all the way into ethics. And there are enough papers about free will in the political freedom subtopic that it won’t push that topic the other way.
A lot of the boundaries here are fuzzy, and this one is fuzzier than most. But I think the division I ended up making looks reasonably plausible.
|History of philosophy||Other history||0.057|
|History of philosophy||Social contract theory||0.058|
|History of philosophy||Heidegger and husserl||0.081|
|History of philosophy||Early modern||0.097|
|History of philosophy||Kant||0.123|
|History of philosophy||Ancient||0.138|
|History of philosophy||Dewey and pragmatism||0.275|
|History of philosophy||Hume||0.315|
|Logic and mathematics||Analytic/synthetic||0.779|
|Logic and mathematics||Universals and particulars||0.780|
|Logic and mathematics||Definitions||0.831|
|Logic and mathematics||Deduction||0.896|
|Logic and mathematics||Verification||0.897|
|Logic and mathematics||Mathematics||0.906|
|Logic and mathematics||Propositions and implications||0.931|
|Logic and mathematics||Vagueness||0.939|
|Logic and mathematics||Sets||0.961|
|Logic and mathematics||Truth||0.974|
This one shouldn’t have been hard, and it wasn’t.
|History of philosophy||Other history||0.034|
|History of philosophy||Social contract theory||0.078|
|History of philosophy||Heidegger and husserl||0.083|
|History of philosophy||Early modern||0.090|
|History of philosophy||Dewey and pragmatism||0.104|
|History of philosophy||Kant||0.121|
|History of philosophy||Ancient||0.127|
|History of philosophy||Hume||0.298|
|Metaphysics||Classical space and time||0.878|
|Metaphysics||Composition and constitution||0.949|
This one could have been harder, but clearly wasn’t either. Note that the model is sure that neither Heidegger nor Dewey are usefully classified with contemporary metaphysics. That’s what I would have said as well, but I was worried it was a biased take on what metaphysics is.
|History of philosophy||Early modern||0.054|
|History of philosophy||Kant||0.083|
|History of philosophy||Other history||0.084|
|History of philosophy||Heidegger and husserl||0.086|
|History of philosophy||Ancient||0.096|
|History of philosophy||Social contract theory||0.186|
|History of philosophy||Hume||0.239|
|History of philosophy||Dewey and pragmatism||0.257|
|Philosophy of language||Meaning and use||0.806|
|Philosophy of language||Language norms||0.883|
|Philosophy of language||Denoting||0.922|
|Philosophy of language||Radical translation||0.922|
|Philosophy of language||Speech acts||0.951|
|Philosophy of language||Belief ascriptions||0.970|
|Philosophy of language||Sense and reference||0.972|
And still this is going smoothly. This is all a bit surprising I think. It’s not like there is much in common between the different parts of history, but the binary sorts still end up grouping them all together.
|History of philosophy||Other history||0.054|
|History of philosophy||Social contract theory||0.074|
|History of philosophy||Early modern||0.106|
|History of philosophy||Kant||0.114|
|History of philosophy||Heidegger and husserl||0.151|
|History of philosophy||Ancient||0.155|
|History of philosophy||Dewey and pragmatism||0.160|
|History of philosophy||Hume||0.185|
|Philosophy of mind||Conceivability arguments||0.359|
|Philosophy of mind||Freud||0.455|
|Philosophy of mind||Self-consciousness||0.567|
|Philosophy of mind||Olp mind||0.641|
|Philosophy of mind||Physicalism||0.727|
|Philosophy of mind||Concepts||0.790|
|Philosophy of mind||Emotions||0.795|
|Philosophy of mind||Psychology||0.822|
|Philosophy of mind||Intention||0.840|
|Philosophy of mind||Minds and machines||0.909|
|Philosophy of mind||Wide content||0.912|
|Philosophy of mind||Color/colour||0.914|
|Philosophy of mind||Perception||0.924|
|Philosophy of mind||Cognitive science||0.945|
And things are still going fairly smoothly. Given how important theories of mind are to some important historical figures, I thought there might be problems here. But it wasn’t.
I don’t know why conceivability arguments ended up seeming so historical. It makes a bit more sense that Freud would feel a bit like a historical topic.
|History of philosophy||Early modern||0.024|
|Philosophy of science||Space and time||0.050|
|History of philosophy||Ancient||0.063|
|History of philosophy||Heidegger and husserl||0.109|
|Philosophy of science||Quantum physics||0.137|
|History of philosophy||Other history||0.195|
|History of philosophy||Kant||0.200|
|Philosophy of science||Chemistry||0.260|
|Philosophy of science||Thermodynamics||0.383|
|History of philosophy||Hume||0.392|
|Philosophy of science||Laws||0.525|
|Philosophy of science||Teleology||0.565|
|History of philosophy||Dewey and pragmatism||0.583|
|History of philosophy||Social contract theory||0.587|
|Philosophy of science||Methodology of science||0.708|
|Philosophy of science||Mechanisms||0.748|
|Philosophy of science||Explanation||0.775|
|Philosophy of science||Theories and realism||0.782|
|Philosophy of science||Grue||0.859|
|Philosophy of science||Models||0.867|
|Philosophy of science||Chance||0.896|
|Philosophy of science||Functions||0.904|
|Philosophy of science||Dna||0.926|
|Philosophy of science||Theory testing||0.953|
|Philosophy of science||Game theory||0.956|
|Philosophy of science||Evolutionary biology||0.960|
And after the last four went so smoothly, this one is a mess. Of all the different boundaries, this is just about the one I was least worried about, and it’s one of the more spectacular failures of my classification tool.
I think what’s happened here is that the binary sort decided that philosophy of physics/philosophy of biology was a more salient dividing line than contemporary philosophy/history of philosophy. So all the physics topics, broadly construed, are at the top of the list, and the biology topics are at the bottom. At least in the parts of history of philosophy that these twelve journals cover, there is a lot more physics-like work than biology-like work, so there is more history at that end. (I’m a bit surprised Ancient ended up at the physics end, I guess.) But both topics from history and topics from philosophy of science that don’t neatly fit on the physics-to-biology spectrum end up clustering in the middle.
Anyway, there weren’t any actual borderline cases here, so it didn’t affect the classification. But we’ll have cause to worry about a similar breakdown in a trickier case soon.
|History of philosophy||Early modern||0.022|
|History of philosophy||Heidegger and husserl||0.033|
|History of philosophy||Other history||0.093|
|History of philosophy||Ancient||0.095|
|Social and political||History and culture||0.161|
|History of philosophy||Dewey and pragmatism||0.180|
|History of philosophy||Kant||0.190|
|Social and political||Life and value||0.200|
|History of philosophy||Hume||0.215|
|Social and political||Marx||0.593|
|History of philosophy||Social contract theory||0.630|
|Social and political||Race||0.659|
|Social and political||Feminism||0.829|
|Social and political||Law||0.844|
|Social and political||Political freedom||0.845|
|Social and political||War||0.880|
|Social and political||Liberal democracy||0.898|
|Social and political||Egalitarianism||0.967|
This one also isn’t neat, and I think that it’s worth going through the topics one by one to double check that we’ve got everything plausibly located.
History and culture really doesn’t feature that many articles about the history of philosophy. It does feature a fair few papers about history, and about the philosophy of history, and I think that’s what confused the model. It’s an easy case, even if the model disagrees.
Life and value is harder. It’s not that it should be in history; there aren’t that many particularly historical papers in it. Remember this topic is something like idealist moral and political philosophy. Now there are some references back to Hegel, and that feels historical. But there are more references to Hegel in idealism. And most of the papers here are trying to put forward first-order philosophical claims, not doing scholarly or exegetical work. I don’t know why this ended up where it did, but it doesn’t feel like a history category.
Marx could have easily been a history topic. But most of the papers here are Marxist analysis of politics and society, not Marx exegesis. I could have gone either way on whether it counted as history or social and political, but I would have tentatively guessed the latter, and it seems the model agrees.
On the other hand, the vast bulk of the papers in social contract theory are clearly history papers. They are history of social and political philosophy, which probably confused the model.
While this is a bit of a mess, the only one that’s really problematic is Life and Value. And I suspect the issue is whether that should be in either of these categories.
|Logic and mathematics||Truth||0.046|
|Logic and mathematics||Deduction||0.049|
|Logic and mathematics||Mathematics||0.088|
|Logic and mathematics||Propositions and implications||0.099|
|Logic and mathematics||Sets||0.117|
|Logic and mathematics||Analytic/synthetic||0.136|
|Logic and mathematics||Definitions||0.139|
|Logic and mathematics||Verification||0.214|
|Logic and mathematics||Universals and particulars||0.269|
|Logic and mathematics||Vagueness||0.443|
|Metaphysics||Classical space and time||0.657|
|Metaphysics||Composition and constitution||0.829|
This was surprisingly straightforward. I would have put universals and particulars in metaphysics, or maybe philosophy of language. (It is largely about predicates, after all.) But if the model wants to include it with logic and mathematics, I’m not going to disagree.
Conversely, I could just as easily have seen modality go with logic as with metaphysics. But it feels more or less natural to include it in metaphysics given the way it is covered in the late twentieth century. (And it is in the first instance a late twentiety-century topic, at least the way this model classifies things.)
This ends up being a case where the binary sort resolves some hard cases, though probably it resolves them in the way I would have done regardless.
|Logic and mathematics||Mathematics||0.039|
|Logic and mathematics||Sets||0.041|
|Logic and mathematics||Deduction||0.122|
|Logic and mathematics||Analytic/synthetic||0.190|
|Logic and mathematics||Universals and particulars||0.197|
|Logic and mathematics||Definitions||0.259|
|Logic and mathematics||Vagueness||0.338|
|Logic and mathematics||Verification||0.377|
|Logic and mathematics||Propositions and implications||0.397|
|Philosophy of language||Radical translation||0.406|
|Logic and mathematics||Truth||0.557|
|Philosophy of language||Denoting||0.644|
|Philosophy of language||Meaning and use||0.672|
|Philosophy of language||Language norms||0.880|
|Philosophy of language||Sense and reference||0.900|
|Philosophy of language||Belief ascriptions||0.952|
|Philosophy of language||Speech acts||0.954|
When I started this project, I was planning to treat logic and language as a single category. I didn’t want the headache of having to think about whether, say, On Denoting was intended as a contribution to philosophy of language or to logic. But then an earlier version of this technique came up with an incredibly clean division of the topics in what I was calling logic and language into two categories, so I split them up. (And added mathematics to the title of the logic category, since it is the mathematics papers that seem to be most paradigmatic.)
And then I tinkered with things and the split wasn’t so clean any more.
But surprisingly the two in the middle that are “out of order” seem like the cleanest cases in the whole list. When looking through the characteristic papers on truth, they are exemplars of what we’d call contemporary work on logic. And if radical translation isn’t a topic in philosophy of language, then I’m not sure what is. So I’m happy to overrule the model on those two cases.
But I’m also happy to have the model decide for me what to say about denoting, and sense and reference, which I could just as easily have classed as language. And from the other direction, I could easily have put analytic/synthetic, definitions and vagueness into language. But I can see why the model made the choices it did, and I suspect my initial judgments were the result of a somewhat partial acquaintance with each of these topics. So I deferred to it on all those cases.
|Logic and mathematics||Sets||0.027|
|Logic and mathematics||Truth||0.036|
|Logic and mathematics||Deduction||0.051|
|Logic and mathematics||Mathematics||0.055|
|Logic and mathematics||Propositions and implications||0.068|
|Logic and mathematics||Vagueness||0.074|
|Logic and mathematics||Definitions||0.179|
|Logic and mathematics||Analytic/synthetic||0.184|
|Logic and mathematics||Verification||0.187|
|Logic and mathematics||Universals and particulars||0.237|
|Philosophy of mind||Conceivability arguments||0.290|
|Philosophy of mind||Freud||0.428|
|Philosophy of mind||Olp mind||0.588|
|Philosophy of mind||Concepts||0.622|
|Philosophy of mind||Minds and machines||0.738|
|Philosophy of mind||Wide content||0.764|
|Philosophy of mind||Cognitive science||0.811|
|Philosophy of mind||Color/colour||0.843|
|Philosophy of mind||Physicalism||0.856|
|Philosophy of mind||Intention||0.874|
|Philosophy of mind||Self-consciousness||0.919|
|Philosophy of mind||Emotions||0.930|
|Philosophy of mind||Perception||0.941|
|Philosophy of mind||Psychology||0.968|
This seems clean enough. The model gets a bit thrown by conceivability arguments, presumably because it associates arguments with logic. But otherwise intuition and the model line up in these cases.
|Philosophy of science||Dna||0.041|
|Philosophy of science||Evolutionary biology||0.041|
|Philosophy of science||Chemistry||0.099|
|Philosophy of science||Thermodynamics||0.103|
|Philosophy of science||Mechanisms||0.112|
|Philosophy of science||Space and time||0.121|
|Philosophy of science||Models||0.141|
|Philosophy of science||Functions||0.160|
|Philosophy of science||Quantum physics||0.199|
|Philosophy of science||Methodology of science||0.225|
|Philosophy of science||Game theory||0.246|
|Philosophy of science||Explanation||0.257|
|Philosophy of science||Teleology||0.392|
|Philosophy of science||Theories and realism||0.419|
|Philosophy of science||Laws||0.522|
|Philosophy of science||Theory testing||0.594|
|Philosophy of science||Chance||0.759|
|Logic and mathematics||Analytic/synthetic||0.842|
|Logic and mathematics||Mathematics||0.842|
|Logic and mathematics||Definitions||0.868|
|Philosophy of science||Grue||0.893|
|Logic and mathematics||Verification||0.898|
|Logic and mathematics||Universals and particulars||0.909|
|Logic and mathematics||Sets||0.932|
|Logic and mathematics||Vagueness||0.937|
|Logic and mathematics||Deduction||0.955|
|Logic and mathematics||Propositions and implications||0.982|
|Logic and mathematics||Truth||0.986|
There are a couple of puzzle cases here, but I think I’m happy to overrule the model in both cases.
One is that the model has a weak preference for putting chance with logic not philosophy of science. This is clearly a mistake—it obviously goes with philosophy of science.
The other is that the model has a strong preference for putting grue with logic. And this is a somewhat more plausible classification. But I think it’s still wrong, for three reasons.
One is that the division here isn’t really between philosophy of science and logic, but between philosophy of biology and logic. All the numbers on the right tell us is where a topic lands on the spectrum from work on, say, the units of selection problem to the semantic paradoxes. It is consistent to say that the grue paradox is more like the semantic paradoxes than it is like the units of selection problem while still saying it is a problem in philosophy of science.
Another is that the model, quite understandably, has a tendency to put any subject matter that includes lots of discussion of conjunctions and disjunctions in with logic. This isn’t entirely wrong, but it is overkill I think. And that’s part of what is driving the classification of grue.
And a third is that the model never really likes splitting up these subtopics, and sets is clearly a logic and mathematics topic, not a philosophy of science topic.
So I’m happy to think there are reasons that the model gets this one case wrong, and also happy that it basically agreed with me on the other twenty-five cases.
|Logic and mathematics||Truth||0.014|
|Logic and mathematics||Propositions and implications||0.019|
|Logic and mathematics||Mathematics||0.020|
|Logic and mathematics||Sets||0.022|
|Logic and mathematics||Vagueness||0.035|
|Logic and mathematics||Deduction||0.040|
|Logic and mathematics||Universals and particulars||0.041|
|Logic and mathematics||Analytic/synthetic||0.053|
|Logic and mathematics||Verification||0.056|
|Logic and mathematics||Definitions||0.061|
|Social and political||Race||0.661|
|Social and political||History and culture||0.812|
|Social and political||War||0.827|
|Social and political||Law||0.844|
|Social and political||Life and value||0.853|
|Social and political||Marx||0.886|
|Social and political||Political freedom||0.915|
|Social and political||Feminism||0.946|
|Social and political||Egalitarianism||0.973|
|Social and political||Liberal democracy||0.973|
A very easy division to make, at least up to 2013. Maybe some of the recent work by folks associated with Pittsburgh and Carnegie Mellon on topics like misinformation, or scientific communities, will complicate this boundary. But up to 2013 it wasn’t complicated at all.
|Philosophy of language||Belief ascriptions||0.053|
|Philosophy of language||Denoting||0.106|
|Philosophy of language||Sense and reference||0.145|
|Philosophy of language||Speech acts||0.171|
|Philosophy of language||Language norms||0.172|
|Metaphysics||Composition and constitution||0.227|
|Philosophy of language||Radical translation||0.444|
|Philosophy of language||Meaning and use||0.752|
|Metaphysics||Classical space and time||0.948|
After a lot of successes, here we have a case where the model is of no use at all. And there is one quite tricky case here: modality.
The model is a bit of a mess. There is no way that meaning and use is really a metaphysics topic, contrary to what the model says. And it is very strange to think it puts composition and constitution with philosophy of language. So I don’t give much weight to what it says about modality. But I do give some weight to it, and so I want to look a bit into whether it is right to put it in metaphysics.
Obviously in philosophy of language we do spend a lot of time talking about modals. And the topic, as you can see from the characteristic articles, has a lot of discussion about conditionals. Because pf this, there are reasons to call it a language topic.
But there are stronger reasons to call it a metaphysics topic. Here are some simple statistics to back this up. Of the 370 articles in it, forty-eight of them (about 13 percent) include the string “world” in the title, and fifty-two of them (about 14 percent) include the string “actual” in the title. These feel like they are metaphysics topics. Twenty-four articles are either by Lewis or have “Lewis” in the title, and these are almost all about modal realism. (Though three are by or about C. I. Lewis, and a couple are about counterfactuals.) Fifty of the articles have “dispos” in the title, and while some of them are kind of about conditionals, most of them are clearly metaphysics papers.
While I could see calling this a language topic, it feels more like metaphysics to me, and that’s how I’ve classified it.
|Metaphysics||Classical space and time||0.059|
|Metaphysics||Composition and constitution||0.070|
|Philosophy of mind||Color/colour||0.105|
|Philosophy of mind||Olp mind||0.227|
|Philosophy of mind||Perception||0.302|
|Philosophy of mind||Physicalism||0.367|
|Philosophy of mind||Conceivability arguments||0.411|
|Philosophy of mind||Concepts||0.611|
|Philosophy of mind||Psychology||0.731|
|Philosophy of mind||Minds and machines||0.781|
|Philosophy of mind||Wide content||0.802|
|Philosophy of mind||Self-consciousness||0.821|
|Philosophy of mind||Emotions||0.881|
|Philosophy of mind||Intention||0.893|
|Philosophy of mind||Freud||0.901|
|Philosophy of mind||Cognitive science||0.946|
Once again the model doesn’t provide particularly useful guidance. And the problem is clear enough - it thinks of metaphysics of mind as more part of metaphysics than of philosophy of mind. And that’s totally understandable, but I think it isn’t true to how philosophy currently thinks of things.
The really tricky case here is color/colour. That could easily go into metaphysics, since a large number of the papers are about what colors really are. But I’ve put it in mind partially because that’s how I conceive of it, and partially to balance out some close calls going in opposite directions. It’s a small topic so it doesn’t make a huge difference to the category statistics, and I think this is the right way to go, but it would be easy to classify it differently.
|Philosophy of science||Space and time||0.054|
|Metaphysics||Classical space and time||0.081|
|Metaphysics||Composition and constitution||0.092|
|Philosophy of science||Quantum physics||0.103|
|Philosophy of science||Thermodynamics||0.301|
|Philosophy of science||Chemistry||0.342|
|Philosophy of science||Laws||0.356|
|Philosophy of science||Teleology||0.590|
|Philosophy of science||Explanation||0.747|
|Philosophy of science||Grue||0.748|
|Philosophy of science||Mechanisms||0.783|
|Philosophy of science||Models||0.803|
|Philosophy of science||Theories and realism||0.810|
|Philosophy of science||Chance||0.848|
|Philosophy of science||Functions||0.869|
|Philosophy of science||Methodology of science||0.905|
|Philosophy of science||Dna||0.936|
|Philosophy of science||Theory testing||0.943|
|Philosophy of science||Game theory||0.945|
|Philosophy of science||Evolutionary biology||0.957|
The binary split the model found here wasn’t between philosophy of science and metaphysics, but between physics and biology. So I have to sort some of the cases by hand.
There ended up being two spacetime subjects. What I’ve called space and time is mostly relativistic space and time, and classical space and time is pre-relativistic. Both of these could have gone just as easily in metaphysics as in philosophy of science. So it seemed like some in-between verdict was called for, and the most natural was to put relativistic work into philosophy of science, and prerelativistic work into metaphysics. The binary sort thinks I should also worry about the classification of quantum physics, but a quick look at the articles in that category—or even at which venues it includes—should persuade you otherwise.
|Social and political||Liberal democracy||0.014|
|Social and political||Egalitarianism||0.030|
|Social and political||Feminism||0.054|
|Social and political||Law||0.080|
|Social and political||Marx||0.093|
|Social and political||War||0.108|
|Social and political||Political freedom||0.109|
|Social and political||History and culture||0.188|
|Social and political||Life and value||0.215|
|Social and political||Race||0.451|
|Metaphysics||Classical space and time||0.979|
|Metaphysics||Composition and constitution||0.979|
After several hard cases in a row, it’s nice to have a simple classification. In 2020, social ontology is a big subject matter, and saying just where metaphysics ends and social and political philosophy begins is hard. In these journals through 2013, it’s not so hard.
It’s kind of striking how not hard it is. composition and constitution would be one of the most tricky borderline subjects if I ran this study on works in contemporary philosophy. But through 2013 it’s just about the easiest example for the model to classify. philosophy sometimes seems like it is spinning in place, but sometimes it changes fast, and hopefully future studies like this one will be able to track some of those changes.
|Philosophy of language||Denoting||0.033|
|Philosophy of language||Sense and reference||0.049|
|Philosophy of language||Belief ascriptions||0.067|
|Philosophy of language||Speech acts||0.076|
|Philosophy of language||Radical translation||0.092|
|Philosophy of language||Language norms||0.144|
|Philosophy of language||Meaning and use||0.180|
|Philosophy of mind||Conceivability arguments||0.310|
|Philosophy of mind||Olp mind||0.491|
|Philosophy of mind||Concepts||0.506|
|Philosophy of mind||Wide content||0.617|
|Philosophy of mind||Freud||0.786|
|Philosophy of mind||Minds and machines||0.815|
|Philosophy of mind||Color/colour||0.859|
|Philosophy of mind||Intention||0.863|
|Philosophy of mind||Physicalism||0.880|
|Philosophy of mind||Cognitive science||0.913|
|Philosophy of mind||Self-consciousness||0.913|
|Philosophy of mind||Perception||0.917|
|Philosophy of mind||Emotions||0.931|
|Philosophy of mind||Psychology||0.975|
This table really surprised me. I think of mind and language as close to a single subject matter. At Michigan one of our most popular big lecture courses is called “Mind and Language”. There is a really important journal called Mind and Language. If I was trying to study contemporary philosophy using data mining and not the history of philosophy, I probably would have included that journal because it is so important. It feels like this should be one of the trickiest boundaries to draw.
And yet, it wasn’t. The only subject that gave the model any pause was the subtopic on conceivability arguments. I would have paused a fair bit over wide content, but the model didn’t really worry about it.
The actual classifications the model makes all look right to me. I’m just surprised it was so definitive.
|Philosophy of science||Quantum physics||0.043|
|Philosophy of science||Space and time||0.044|
|Philosophy of science||Thermodynamics||0.051|
|Philosophy of science||Evolutionary biology||0.052|
|Philosophy of science||Dna||0.056|
|Philosophy of science||Models||0.067|
|Philosophy of science||Chemistry||0.074|
|Philosophy of science||Theory testing||0.114|
|Philosophy of science||Chance||0.126|
|Philosophy of science||Explanation||0.196|
|Philosophy of science||Mechanisms||0.231|
|Philosophy of science||Theories and realism||0.235|
|Philosophy of science||Functions||0.257|
|Philosophy of science||Methodology of science||0.266|
|Philosophy of science||Laws||0.286|
|Philosophy of science||Game theory||0.302|
|Philosophy of science||Grue||0.408|
|Philosophy of science||Teleology||0.675|
|Philosophy of language||Radical translation||0.903|
|Philosophy of language||Language norms||0.961|
|Philosophy of language||Meaning and use||0.967|
|Philosophy of language||Speech acts||0.975|
|Philosophy of language||Sense and reference||0.979|
|Philosophy of language||Denoting||0.988|
|Philosophy of language||Belief ascriptions||0.991|
This one, on the other hand, should have been easier. And it mostly was. Once again the model struggled with a subtopic, in this case teleology, but otherwise I don’t see much here to quibble about. Maybe looking forward game theory will become more of a borderline case, but for the most part these are separate disciplines.
|Social and political||Egalitarianism||0.022|
|Social and political||Liberal democracy||0.024|
|Social and political||Marx||0.071|
|Social and political||Life and value||0.076|
|Social and political||Political freedom||0.082|
|Social and political||Feminism||0.083|
|Social and political||History and culture||0.119|
|Social and political||Law||0.127|
|Social and political||War||0.164|
|Social and political||Race||0.225|
|Philosophy of language||Language norms||0.850|
|Philosophy of language||Meaning and use||0.932|
|Philosophy of language||Speech acts||0.949|
|Philosophy of language||Radical translation||0.968|
|Philosophy of language||Belief ascriptions||0.981|
|Philosophy of language||Sense and reference||0.986|
|Philosophy of language||Denoting||0.990|
And this really brings out the difference between philosophy in 2020 and philosophy in these journals up to 2013. Social and political philosophy of language is one of the fastest growing fields in philosophy. The literature on slurs alone is big enough to be a subject in its own right. But add in work on silencing, on propaganda, on trust, lying and deception, and so on, and you have a huge body of work that should be hard to clearly sort into one of these categories. And in the journals up to 2013, we see virtually none of it. Let’s look back at this when we have some more data and see how populated this boundary gets.
|Philosophy of science||Quantum physics||0.045|
|Philosophy of science||Thermodynamics||0.048|
|Philosophy of science||Models||0.050|
|Philosophy of science||Evolutionary biology||0.063|
|Philosophy of science||Dna||0.064|
|Philosophy of science||Space and time||0.076|
|Philosophy of science||Theories and realism||0.123|
|Philosophy of science||Theory testing||0.126|
|Philosophy of science||Chance||0.139|
|Philosophy of science||Chemistry||0.160|
|Philosophy of science||Explanation||0.191|
|Philosophy of science||Methodology of science||0.259|
|Philosophy of science||Grue||0.262|
|Philosophy of science||Functions||0.267|
|Philosophy of science||Laws||0.363|
|Philosophy of science||Mechanisms||0.485|
|Philosophy of science||Game theory||0.516|
|Philosophy of mind||Freud||0.587|
|Philosophy of mind||Cognitive science||0.589|
|Philosophy of science||Teleology||0.754|
|Philosophy of mind||Minds and machines||0.781|
|Philosophy of mind||Conceivability arguments||0.798|
|Philosophy of mind||Physicalism||0.829|
|Philosophy of mind||Concepts||0.879|
|Philosophy of mind||Wide content||0.892|
|Philosophy of mind||Olp mind||0.922|
|Philosophy of mind||Psychology||0.925|
|Philosophy of mind||Color/colour||0.929|
|Philosophy of mind||Intention||0.931|
|Philosophy of mind||Self-consciousness||0.963|
|Philosophy of mind||Emotions||0.974|
|Philosophy of mind||Perception||0.975|
This came out a little neater than I expected. It got the subtopic on teleology wrong again, but otherwise it looks pretty good. Given the amount of scientific work that turns up in philosophy of mind, especially in the nineteenth and twenty-first centuries, I thought this would confuse the model more than it did.
|Social and political||Liberal democracy||0.019|
|Social and political||Egalitarianism||0.022|
|Social and political||Feminism||0.098|
|Social and political||Political freedom||0.099|
|Social and political||Law||0.100|
|Social and political||War||0.120|
|Social and political||Marx||0.127|
|Social and political||History and culture||0.216|
|Social and political||Race||0.305|
|Social and political||Life and value||0.340|
|Philosophy of mind||Freud||0.453|
|Philosophy of mind||Conceivability arguments||0.712|
|Philosophy of mind||Intention||0.744|
|Philosophy of mind||Emotions||0.815|
|Philosophy of mind||Self-consciousness||0.863|
|Philosophy of mind||Concepts||0.933|
|Philosophy of mind||Minds and machines||0.946|
|Philosophy of mind||Cognitive science||0.950|
|Philosophy of mind||Psychology||0.952|
|Philosophy of mind||Olp mind||0.959|
|Philosophy of mind||Physicalism||0.975|
|Philosophy of mind||Wide content||0.984|
|Philosophy of mind||Color/colour||0.987|
|Philosophy of mind||Perception||0.989|
Not many surprises here. The model doesn’t quite know what to do with philosophical articles about Freud. I don’t know what to do with philosophical articles about Freud. But otherwise it doesn’t see a great deal of overlap. Maybe if we ran this study forward some of the recent work on implicit bias would confound it a little.
|Philosophy of science||Quantum physics||0.007|
|Philosophy of science||Space and time||0.008|
|Philosophy of science||Dna||0.016|
|Philosophy of science||Thermodynamics||0.034|
|Philosophy of science||Theories and realism||0.042|
|Philosophy of science||Chemistry||0.044|
|Philosophy of science||Models||0.047|
|Philosophy of science||Explanation||0.077|
|Philosophy of science||Evolutionary biology||0.087|
|Philosophy of science||Chance||0.098|
|Philosophy of science||Theory testing||0.100|
|Philosophy of science||Laws||0.116|
|Philosophy of science||Grue||0.129|
|Philosophy of science||Functions||0.130|
|Philosophy of science||Mechanisms||0.219|
|Philosophy of science||Methodology of science||0.290|
|Philosophy of science||Teleology||0.457|
|Social and political||Race||0.558|
|Social and political||History and culture||0.678|
|Philosophy of science||Game theory||0.690|
|Social and political||Life and value||0.790|
|Social and political||Marx||0.834|
|Social and political||Law||0.911|
|Social and political||War||0.913|
|Social and political||Political freedom||0.942|
|Social and political||Feminism||0.943|
|Social and political||Egalitarianism||0.971|
|Social and political||Liberal democracy||0.986|
The only tricky case here is game theory. Since game theory is in its nature about the study of social groups, it isn’t surprising that the model wans to put it with science and political philosophy. But looking at the particular articles in that topic, which largely focus on evolutionary game theory, I think it’s a much better fit with the other philosophy of biology articles in philosophy of science.
I’ve spent a bit of time on going over all thirty-six of these boundaries for two reasons.
One is that the classifications of topics into categories is crucial for generating the category graphs. And there were a lot of choices to be made in generating that classification that could have gone either way, and I wanted to lay out a bit why I made the choices that I did.
The other is that looking at these boundaries is a pretty interesting perspective on how the future might fail to resemble the past. One tried and trusted way to make philosophical progress is to start with two areas that aren’t in a lot of contact, and see what happens when you use the tools and methods of each to look at the questions of the other. We’ve gone through a couple of decades of doing that with ethics and epistemology, and I think the results have been very rewarding. The period since this study ends has seen an explosion of work intersecting science and political philosophy with any number of the other fields here: epistemology, metaphysics and philosophy of language being particularly important.
In general, thinking about where the sharp boundaries are in a study like this, and about what kind of work would make those boundaries less sharp, is a useful way to think about where more work could be usefully done.