Monday, October 16, 2017

Historical Period Designations

I was somewhat amused by this post by Scott Alexander on the Dark Ages. The primary reason is that I am somewhat amused by people talking about the Dark Ages in general; when people talk about this period of "profound economic and intellectual decline and stagnation" I always say, "For whom?" The exact period in question, from 500 to 1000, was a golden age of expansion and increasing prosperity -- for my Scandinavian ancestors. The early part of the period is the Germanic Iron Age: significant influx of gold and metals, consolidation of Scandinavian trade, increasing artistic sophistication, the Lombards finishing their long migration into Italy. The later part of the period is the Viking Age: increasing dominance of trade routes, increasing possession of England and Scotland, the rise of the Varangians, the rise of the Normans, the rise of the Three Kingdoms, invention of the Althing in Iceland, far-flung settlements. There were setbacks, such as Ireland and Andalusia, and various civil wars, but, by and large, it seems to have been an age of improvement and progress for all Scandinavians. Splendid, splendid days; we should call them the Awesome Ages.

But, jokes and amusement aside, what interests me more than the question of whether the Dark Ages were dark is the error that is commonly made when people discuss historical topics like this, namely, treating period designations as part of the data. It's interesting and worth thinking about on its own. We get a pretty clear example of the mistake here:

The period from about 500 to about 1000 in Christian Western Europe was marked by profound economic and intellectual decline and stagnation relative to the periods that came before and after it. This is incompatible with the “no such thing as the Dark Ages” claim except by a bunch of tortured logic, isolated demands for rigor, and historical ignorance.

And another example here:

Every other historical age name is instantly understood by everyone to refer to both a time and a place. The only time anyone ever gives anybody else grief over this is when they talk about the Dark Ages. This is an isolated demand for rigor. And if this is really your true objection, let’s just agree to call it the Western European Dark Ages, as long as we can also agree it existed and was bad.

Both of these express a misunderstanding about how period designations work in any kind of historical field; they treat the designated period and its label as if it were a natural feature of the data whose use requires no justification. Legitimate period designations are tools, however. They are not built into the data, but arise from the confluence of three factors:

(1) identifiable events, coherent enough to be more intelligible if classified together than they are if treated separately;
(2) the need to state what you are doing in shorthand, even if the way of doing it is primarily practical convenience;
(3) constraints arising from academic life itself.

All three are always operative, but not all are equally important for particular naming practices. Thus, for instance, Scott assumes, just before the second passage quoted above, that we name the Warring States period as we do because there were a lot of warring states, when in reality we call it the Warring States period because it is, more or less, the period covered by the Record of the Warring States. To be sure, there were a lot of warring states in the period, but there were also a lot of springs and autumns in the Spring and Autumn period; it's not the reason for the name, and if the Record of the Warring States had instead been called the Honey Milk Book, we'd likely be calling the era the Honey Milk period. Thus this is mostly an example of (2), not (1)-- one can give a (1)-ish account for the designation in terms of the breakdown of the Zhou dynasty, but doing so will inevitably result in complications that don't arise with a (2)-ish use -- for instance, some things in the early Warring States period might make more sense grouped with Spring and Autumn events than late Warring States events, and some late Warring States events might make more sense grouped as part of the rise of the Qin dynasty rather than with early Warring States events. And even when one takes it as a whole, there would be perfectly legitimate questions about whether it was misleading if lots and lots of people, assuming with Scott that the name is primarily descriptive rather than primarily referential, started making value judgments about the period on the basis of what it was called. We see the same conflation in Scott's argument when he says, "Historical periods get their names from random individuals reflecting on them; the names catch on if people agree that they fit." Sometimes. They also sometimes catch on because people need a communicative shorthand enough that it will do even if it is very flawed and does not fit very well at all, and all the examples Scott identifies are clearly cases where the label was proposed not because it 'fit' but because some people started using it as shorthand description and other people started using it because they too needed a shorthand description, and, lo! here there already was one.

A good example of label use that mostly has a (3)-ish account is the notion of a 'Middle Ages', which as its very name implies is a omnium gatherum for what happens between the Ancient and the Modern. Talk to medieval historians, and most would be happy to get rid of the designation, which is not particularly great for conveying any kind of information, and would prefer to replace it with half a dozen different ones that break the whole thousand-or-so years into more natural clumps of events. But in terms of how things are taught, funding that is provided, the need of scholars to engage in cooperative endeavors combined with the limitations that arise from having too few scholars to band together in such endeavors, the slow change of widespread naming conventions, it's just inevitable that the term keeps being used, even though it isn't particularly useful for historical research, and has some genuine disadvantages.

Thus period designations are not what historians study; they are classifications designed to facilitate that study, either because they are reasonably natural or informative given the evidence, or because they are practically convenient, or because they facilitate the smooth running of academic life. Thus it's already something of an illegitimate question to ask "Were the Dark Ages really dark?"; the real questions are, "Is the term 'Dark Ages' an appropriate (1)-ish term even to begin with?" and "Does the 'Dark Ages' serve as a shorthand whose practical convenience outweighs its potential to be misunderstood?" There is no such thing 'the Dark Age' in the historical data; 'the Dark Age' is a label that we use to group the historical data, if it is reasonable or useful.

Thus, to take another example, in my own field, one can reasonably argue that the Enlightenment was 'not a thing'; it's not, as far as I can tell, a particularly common position, but it is an intelligible one that you do occasionally find. But, someone might say, there are literally people in the period referring to their period as a period of Enlightenment! Yes, but:

(A) There were others who didn't, and there were lots of things going on that had little to do with these thinkers. Since the people who used the term (and have since used the term) often had a very open agenda about using it, one can reasonably question whether they were right in understanding their own times, or whether using the term buys too much into their particular valuations, to the detriment of understanding other things.
(B) It's pretty clear that the term 'Enlightenment' is often confusing to people who are not specialists -- they don't often distinguish it very well from other periods, with the result that 'Enlightenment' is often used in public for things that really only arose after the period you're trying to discuss.
(C) Even if one uses the 'Enlightenment' designation, it often makes more sense to think of it as several different things rather than a single thing -- that is, instead of 'the Enlightenment' to talk of the French Enlightenment, the Scottish Enlightenment, the American Enlightenment, the Swiss Enlightenment, and so forth. That is to say, whether you are in 'the Enlightenment period' depends in part on which area of the world you are taking as your reference point, and isn't always very useful taken as a neutral term covering everything. Thinking of this all as 'the Enlightenment' is much later. (This and (A) are probably the most widely accepted reasons among specialists for not putting much weight on 'Enlightenment' as a designation.)
(D) It's always a reasonable question, even if the designation is useful, whether there is a better one.

None of this is determinative of itself; in practice, people who would deny that the Enlightenment is a thing in the (1)-ish sense would probably often still use it in the (3)-ish sense, since it is a word that non-specialists recognize and that, because it is usually associated with positive value judgments, would usually not cause any problems for funding or the like. With the Dark Ages, it's unclear what analogous (3)-ish value would be operative; people might fund Enlightenment Studies, but good luck getting public interest and funding for Dark Ages Studies. They might also use it in a (2)-ish sense -- i.e., 'I study the Enlightenment, understood in the sense of the things that were happening in intellectual matters that Kant probably had in mind when he talked about Aufklärung'), although, because of (B), they might not. But if they are denying that it is a 'thing', it's a (1)-ish question -- and, indeed, rigor is the order of the day then.

In the case of the Dark Ages, people who deny that it is a 'thing' are typically denying that it is a natural classification of historical events and that it is a shorthand that sufficiently avoids misleading people. Scott considers both of these but, as noted before, does not adequately distinguish them. Worries about value judgments are typically (2)-ish problems:

So I assume you also raise a fuss whenever someone talks about Alexander the Great? The Golden Age of Athens? The Five Good Emperors? The Enlightenment? Ivan the Terrible? The Belle Époque? I S O L A T E D . D E M A N D . F O R . R I G O R.

But this line of argument is just evidence that Scott doesn't spend an extensive amount of time talking with historians, because historians worry about this kind of thing all the time, particularly when referential uses capture the public imagination as if they were descriptive. 'The Enlightenment' is indeed the sort of label that can worrisomely introduce value-based prejudices that distort serious scholarship, and it's entirely possible to worry about it, and some people do, as I noted above. Historians often question the epithets and descriptions attached to figures in popular history. They will still use them, however, if they don't see any reason to think that confusion between the descriptive reading and the referential reading are misleading people -- particularly when they have (3)-ish incentives for doing so.

But Scott also muddles up with this (1)-ish problems about whether the Dark Ages were really bad -- which, first of all, assumes that there is already particular reason to treat clump these centuries together under a unified label rather than break them up or portion off part to the period before and part to the period after. If one has such a reason, that's a (1)-ish ground for the label. And second it raises the question of whether the label appropriately conveys what is going on even if there is reason to treat it as unified. Certainly there were bad things that happened; the question of significance is whether it was enough of a single bad network of things to warrant a single label, or just a bunch of bad clumps. And even if one argues the former, talking about 'the Dark Ages' makes it sound -- well, darker than everything else, since you are singling it out in particular as dark; and that is a comparative judgment that requires not only looking at the bad (as Scott does) but the good. This is because the label does not exist on its own but is an element within a labeling system, and its place in that system needs to be considered in a comparative matter like this.

Consider an example. The twentieth century saw something like 160 million people die from war, and probably as many in non-war killings by dictatorial regimes; it saw a nation drop atomic weapons on another nation; it saw major plague outbreaks and any number of other bad things. The end of it sees the collapse of a number of previously important institutions and the loss of cultural customs around the world, serious collapses of popular trust in governments, religious institutions, and scientific inquiry. But if in the future it gets called the Bad Century, one would have to look at the good of the twentieth century as well as the bad before one could properly determine whether the label was reasonable; just as one would also have to consider the question of whether the accumulation of badness was really obtained by jumping around and treating all these bad things as if they characterized the century rather than just this set of events at this particular point in this particular population; and one would also need to consider whether 'twentieth century' were too arbitrary a designation and whether its events would actually be better understood if split up in different ways, even though people in the twentieth century did tend to think of the twentieth century as a block. (Thus, to take just one for-instance, when Scott talks about population decline, he fails to consider the question of whether his data actually suggest 200 to 600 should be seen as cohering better as a Late Imperial period, or an Imperial Decline period, and then 600 to 1000 as a Recovery period. Assuming we should regard this as a single period, which is part of what is being questioned when people deny that the Dark Ages is a 'thing', why the pessimistic reading in which the entire era is blamed for a problem that obviously started before it supposedly began, since Scott takes the Dark Ages to start circa 500? The only reason is because there is already a label 'Dark Ages' and the pessimistic reading fits it better. This same thing actually happens several times in the argument: he is conflating the use of the label to interpret the data and the derivation of it from the data, so that his defense of the label as a legitimate one depends on using the label in the first place.) The use of the label is not self-justified just because it is used; if we are reading it in a (1)-ish way, we have to establish that it actually designates something non-arbitrary that is not better classified in a different way, and that the label is a reasonable label when put into the entire classification scheme composed by historical period designations.

Again, the point is not the question of the Dark Ages, which is an extraordinarily complicated historical question, but instead that of how historical period designations work. As Whewell pointed out, classifications are not trivial issues because they are one of the ways we store discoveries and they are instruments we use to organize evidence and research. The labels used for historical periods are classifications. Like other classifications, their use may be due to the fact that they converge on a natural classification, or because they are useful enough even if artificial to make research easier, or because they allow for the smooth functioning and administration of institutions and organizations that do the research. As with other classifications, one may criticize them on any or all of these grounds. But they are distinct.