5 ms·

Interesting discussion here. I don't think you're representing the author's point correctly, though, given the larger context of the work he's doing and current

Interesting discussion here. I don't think you're representing the author's point correctly, though, given the larger context of the work he's doing and current debates in the historical field. Here's what he says: "I suppose the point of this post is to articulate my growing concern that we are so damn good at coming up with post-facto historical explanations to contextualize any given observation, that we are particularly susceptible to confabulating these post-facto rationalizations with the idea that we somehow knew the results of this quantitative work already." The implication here, I think, is simply that historians should be open to multiple ways of testing explanations (i.e., combining qualitative and quantitative approaches, if those are available). And in the larger context of debates about digital humanities, he's swatting down the argument that DH simply tells us what we already knew, because if done right it can be used to add a new perspective on existing arguments. The point isn't that history should be science but that "scientific"/quantitative approaches can be used to test the validity of historical claims based on more traditional historical work, like finding manuscripts in archives, doing a close reading of a key text, tracing correspondence networks, tracing the material history of a painting or object, etc. I agree with him. Edited to add: also need to be careful about confusing prediction with explanation. Historians don't predict much, but we certainly try to explain things a lot! I read this post as being exclusively about the latter.

pron 12y ago
I agree with everything you say (although I read this article differently), but historians should be careful to steer away from Hayek's "scientism". History's underlying mathematical assumption -- an assumption that is almost surely correct -- is that society is a complex system, and therefore cannot be subject to predictions (or very few, qualitative predictions) in the long run. If historians feel quantitative analysis gives their explanations any sort of definitiveness (and "proving" the inevitability of past events is exactly the same as prediction), they will fall into the same trap as economists. A different perspective -- absolutely; a quantitative theory -- no. In fact, quantitative theory, at least in a sense similar to physics, is nearly impossible even in subjects that are much closer to physics than history. The simplest non-linear differential equations defy quantitative study, and history (and even economics and biology) are nowhere near simple. However, it is possible that even in a complex system, there should arise some temporary structures that can be described as linear processes, and therefore subject to some quantitative theory, but it will be short lived. Complex systems undergo phase-changes that tend to break and reform any structure. On the other hand, it can be argued that those phase changes are rare, and that human society is self-stabilizing on some global scale, so that the system -- at least for the moment -- is stable, and some quantitative analysis may apply. But I'm pretty sure no one would classify any specific society (i.e. a small subset of humankind) as stable.
- benbreen 12y ago
  Agreed. A big part of why I enjoy thinking about historical problems so much is that it's virtually impossible to supply a definitive explanation for anything (we're still arguing over the Roman empire's demise one and a half millennia on). I think the author's point - or at least the stance I take, which I suspect he aligns generally with too based on his work - is that quantitive data can only be used in historical arguments in what is effectively a qualitative way, i.e. as part of a larger, holistic assessment based on many perspectives rather than as a definitive "proof" of anything. But it's an exciting addition because historians typically haven't taken this approach at all prior to the last decade, excepting some forays into punchcard computing in the 60s and 70s by people like Lawrence Stone and the Annalists. Incidentally, I'd point to Pinker's "Better Angels of Our Nature" book as an example of the risks of putting too much stock in a scientific/econometric approach to history. I actually think his argument is (in broad outlines) convincing, but the way he marshals his evidence drives academic historians crazy because it ignores that history is a complex system, as you say. (i.e., he does things like saying "x tribe in New Guinea is a 'stone age' tribe and according to one or two studies has a high murder rate, therefore all humans in 'the stone age' had a murder rate comparable to that tribe.")
- tjradcliffe 12y ago
  Predictions can be sideways as well as forward. If one gives an explanatory account of some phenomenon, the existential claims made as part of that explanation should in general have implications for other parts of the same place and time, so by asking if those implications are plausibly supported by what we know of that place and time is a way of testing the idea that "phenomenon X is explained by Y" while still being in a position where it is impossible to make predictions about the future, or other places and times.
  - pron 12y ago
    Yes, the problem is that complex systems are complex "sideways" as well (i.e. in all dimensions). Sorry, but that's the nature of non-linear dynamics. They defy quantitative study. That's not to say that it isn't interesting or worthwhile to come up with comparative analysis to explain, say, the "European Miracle", but it can never have the same definitiveness as Newton's laws.
tjradcliffe 12y ago
Anyone concerned with this stuff should read: "Everything is Obvious (once you know the answer)" by Duncan Watts. http://www.amazon.com/Everything-Obvious-Common-Sense-Fails/dp/0307951790/ref=sr_1_1?ie=UTF8&qid=1427159236&sr=8-1&keywords=everything+is+obvious http://www.amazon.com/Everything-Obvious-Common-Sense-Fails/... It isn't a great book, but it does clearly identify the difference between explanatory and predictive accounts, and makes a strong case that explanatory, after-the-fact, accounts are simply noise. The exercise the author engages in here, of "making sense" of two contradictory datasets can be done with almost any phenomenon. The more systematic and inclusive data gathering is, the more strongly any explanation is going to be tied to the phenomenon it is going to explain, and historians--like any science--should always be asking themselves, "How can I test this explanatory idea? If it is true, then what else ought to be true?" Otherwise, when you "explain things a lot" you are doing nothing but generating noise. The financial press provides strong evidence for this: tens of thousands of words of "explanation" every single day, and not a single person made rich by any of the predictions the same people make. When you can explain everything and predict nothing, you had better be able to test your explanations by indirect means if you want to be taken more seriously than a fabulist. The latter part is important: historians and others like them expect us to take their pontifications seriously. But we know the untested explanation is almost certainly nothing but a confabulation. So why should we take anything historians say seriously? The history of science--which I define inclusively as "publicly testing ideas by systematic observation, controlled experiment and Bayesian inference"--is one of realizing we've been doing it wrong, that ideas we believed and took seriously were actually rather silly. Blood circulates. Geese don't come from barnacles. Bad air or moral turpitude don't cause plague. And so on. If historians are starting to ask, "Are we doing this wrong?" that's a good thing, and based on my understanding of what science is and how it works I predict that over the next few decades the "digital humanities" will make a lot of traditional beliefs look untenably foolish, and there will be a general shift of the field toward Bayesian methods that will alienate and annoy a great many people whose freedom to confabulate will be greatly curtailed. It will change the face of the humanities to the extent that they will have as little to do with their historical roots as modern psychology has to do with the story-tellers of the early 20th century who gave it its start.
- pron 12y ago
mdlincoln 12y ago
Yes, I am guilty of writing the post for an audience already largely familiar with the context. I should probably add that the types of "explanations" I put forward in this post are actually not of central concern to me - certainly not explanations derived solely from parsing quantitative results. I'm far more interested in the descriptive evidence this kind of measurement can provide. It can give wider context to what tends to be a very case-study-centric discipline (e.g. oh, this guy happened to work a lot with Italian publishers in this period? We didn't realize it before just looking at 5-10 artists per article/monograph, but actually that is quite exceptional/normal for this period...) Then again, proposing these kinds of explanations is also something of a disciplinary norm, for better or worse.

pron 12y ago

I agree with everything you say (although I read this article differently), but historians should be careful to steer away from Hayek's "scientism". History's underlying mathematical assumption -- an assumption that is almost surely correct -- is that society is a complex system, and therefore cannot be subject to predictions (or very few, qualitative predictions) in the long run. If historians feel quantitative analysis gives their explanations any sort of definitiveness (and "proving" the inevitability of past events is exactly the same as prediction), they will fall into the same trap as economists. A different perspective -- absolutely; a quantitative theory -- no. In fact, quantitative theory, at least in a sense similar to physics, is nearly impossible even in subjects that are much closer to physics than history. The simplest non-linear differential equations defy quantitative study, and history (and even economics and biology) are nowhere near simple. However, it is possible that even in a complex system, there should arise some temporary structures that can be described as linear processes, and therefore subject to some quantitative theory, but it will be short lived. Complex systems undergo phase-changes that tend to break and reform any structure. On the other hand, it can be argued that those phase changes are rare, and that human society is self-stabilizing on some global scale, so that the system -- at least for the moment -- is stable, and some quantitative analysis may apply. But I'm pretty sure no one would classify any specific society (i.e. a small subset of humankind) as stable.

benbreen 12y ago
Agreed. A big part of why I enjoy thinking about historical problems so much is that it's virtually impossible to supply a definitive explanation for anything (we're still arguing over the Roman empire's demise one and a half millennia on). I think the author's point - or at least the stance I take, which I suspect he aligns generally with too based on his work - is that quantitive data can only be used in historical arguments in what is effectively a qualitative way, i.e. as part of a larger, holistic assessment based on many perspectives rather than as a definitive "proof" of anything. But it's an exciting addition because historians typically haven't taken this approach at all prior to the last decade, excepting some forays into punchcard computing in the 60s and 70s by people like Lawrence Stone and the Annalists. Incidentally, I'd point to Pinker's "Better Angels of Our Nature" book as an example of the risks of putting too much stock in a scientific/econometric approach to history. I actually think his argument is (in broad outlines) convincing, but the way he marshals his evidence drives academic historians crazy because it ignores that history is a complex system, as you say. (i.e., he does things like saying "x tribe in New Guinea is a 'stone age' tribe and according to one or two studies has a high murder rate, therefore all humans in 'the stone age' had a murder rate comparable to that tribe.")
tjradcliffe 12y ago
Predictions can be sideways as well as forward. If one gives an explanatory account of some phenomenon, the existential claims made as part of that explanation should in general have implications for other parts of the same place and time, so by asking if those implications are plausibly supported by what we know of that place and time is a way of testing the idea that "phenomenon X is explained by Y" while still being in a position where it is impossible to make predictions about the future, or other places and times.
- pron 12y ago
  Yes, the problem is that complex systems are complex "sideways" as well (i.e. in all dimensions). Sorry, but that's the nature of non-linear dynamics. They defy quantitative study. That's not to say that it isn't interesting or worthwhile to come up with comparative analysis to explain, say, the "European Miracle", but it can never have the same definitiveness as Newton's laws.

tjradcliffe 12y ago

Anyone concerned with this stuff should read: "Everything is Obvious (once you know the answer)" by Duncan Watts. http://www.amazon.com/Everything-Obvious-Common-Sense-Fails/dp/0307951790/ref=sr_1_1?ie=UTF8&qid=1427159236&sr=8-1&keywords=everything+is+obvious http://www.amazon.com/Everything-Obvious-Common-Sense-Fails/... It isn't a great book, but it does clearly identify the difference between explanatory and predictive accounts, and makes a strong case that explanatory, after-the-fact, accounts are simply noise. The exercise the author engages in here, of "making sense" of two contradictory datasets can be done with almost any phenomenon. The more systematic and inclusive data gathering is, the more strongly any explanation is going to be tied to the phenomenon it is going to explain, and historians--like any science--should always be asking themselves, "How can I test this explanatory idea? If it is true, then what else ought to be true?" Otherwise, when you "explain things a lot" you are doing nothing but generating noise. The financial press provides strong evidence for this: tens of thousands of words of "explanation" every single day, and not a single person made rich by any of the predictions the same people make. When you can explain everything and predict nothing, you had better be able to test your explanations by indirect means if you want to be taken more seriously than a fabulist. The latter part is important: historians and others like them expect us to take their pontifications seriously. But we know the untested explanation is almost certainly nothing but a confabulation. So why should we take anything historians say seriously? The history of science--which I define inclusively as "publicly testing ideas by systematic observation, controlled experiment and Bayesian inference"--is one of realizing we've been doing it wrong, that ideas we believed and took seriously were actually rather silly. Blood circulates. Geese don't come from barnacles. Bad air or moral turpitude don't cause plague. And so on. If historians are starting to ask, "Are we doing this wrong?" that's a good thing, and based on my understanding of what science is and how it works I predict that over the next few decades the "digital humanities" will make a lot of traditional beliefs look untenably foolish, and there will be a general shift of the field toward Bayesian methods that will alienate and annoy a great many people whose freedom to confabulate will be greatly curtailed. It will change the face of the humanities to the extent that they will have as little to do with their historical roots as modern psychology has to do with the story-tellers of the early 20th century who gave it its start.

pron 12y ago

mdlincoln 12y ago

Yes, I am guilty of writing the post for an audience already largely familiar with the context. I should probably add that the types of "explanations" I put forward in this post are actually not of central concern to me - certainly not explanations derived solely from parsing quantitative results. I'm far more interested in the descriptive evidence this kind of measurement can provide. It can give wider context to what tends to be a very case-study-centric discipline (e.g. oh, this guy happened to work a lot with Italian publishers in this period? We didn't realize it before just looking at 5-10 artists per article/monograph, but actually that is quite exceptional/normal for this period...) Then again, proposing these kinds of explanations is also something of a disciplinary norm, for better or worse.