There are three major institutional incentives that I see which negatively affect the social sciences. They are (a) institutional hierarchy; (b) a factitious division of intellectual labor; and (c) publish or perish.
I’ll explain how I think these operate. But let’s begin with a few observations that spurred me to analyze the institutional conditions of the social sciences.
Below are several recent trends in social sciences generally, and cultural studies and the humanities in particular, as I see them. 
Loss of rational argument
Especially in the humanities, one notices a strong tendency in the last half-century to abandon the canons of rational inquiry and argument. Anyone who’s read Hegel or Heidegger knows that there is a vein of philosophy that does not take rational argument very seriously. But contemporary cultural studies has come in recent decades to amplify this trend.
Within this context, one finds that theses and claims are not presented clearly, if they are present at all. Evidence is not adduced in any systematic way; often a mere token reference is left to stand in for a range of potential evidence.
There is no null hypothesis (what the case would look like if the thesis/claim is incorrect); no sense of claims even being related to the notion of “correct” or “incorrect.”
Often claims are presented in a non-falsifiable manner, meaning they can never be right or wrong, since you can never prove them in either direction. The common method here is simply to make a sufficiently vague claim, or include vague terms within the claim, such that the claim is inherently untestable.
Compare the standards of argument present in many of today’s trendy journals, especially in the culture studies scene, to intellectual argumentation and rational inquiry since Bacon, culminating, in the social sciences in figures like John Stuart Mill or Bertrand Russell. When you read Descartes, there are clear claims being made, there is a path of argumentation that is clearly presented, with evidence adduced for each claims. There is, other words, a way to logically evaluate the argument and the conclusion.
Compare this to any text by, say, Jacques Derrida or Jacques Lacan. It is often difficult to even locate the claims being made. One is often forced to invent an implicit claim just to get any sense out of the text. When there is any detectable claim, it is formulated in such vague terms that there is no way to evaluate it. When there is evidence provided, it is not a representative sample, but a single example: cherry-picked for the purposes of the claim, which means it proves nothing. Often, an entire text by such authors reduces to a pile of confused thinking, the sort of thing most teachers, if they received such a paper from their students, would mark as: “poor argument, unclear presentation, requires revision.”
No defined, common problem or object
The hard sciences have the luxury of finding ready-made objects of inquiry before their eyes, at least in the beginning. So: explain the sun; explain the movement of the stars; explain why things fall; explain colors; explain the human body; explain this rock.
Social sciences do not have the luxury of such ready made objects, since their field is human affairs. Objects of inquiry don’t fall out of the sky, so they have to be carved out by a definition: explain X, where X is something we have defined as a coherent concept and an existing object. Examples: an economy; a culture; a psyche; a behavior; a political arrangement; a work of literature. Here though is a first stumbling block.
The sciences realized long ago that appearances are deceptive, and many appearances which at first seem to be coherent problems or concepts, in fact turn out to need redefining. So the problem of “light” and “color” becomes in fact a problem of things called photons and ocular systems; the problem of rocks becomes a problem of molecular behavior, planet formation, and tectonics; “things falling” turns out to be a problem of occult forces popularly termed “gravity.” The redefining of problems was a normal result of defining the object with precision. As you describe the phenomena, it turns out that the common sense terms, perfectly valid for practical purposes, are useless for scientific purposes, because they are imprecise. When you start defining things with precision, you end up needing totally new concepts, terms, and thus redefining your problems.
The social sciences generally seem to be quite unaware of these matters. You often see terms like “democracy” or “the sublime” or “Europe” or “Christianity” used as if they represented precisely-defined entities actually present and observable in the world. Then important claims are made using such extremely vague concepts. A more careful approach would either admit that only limited claims can be made with the common sense terms, or else define the term in a very precise manner, in order to make it vulnerable to observation (i.e., the criterion of falsifiability). (Some strains of the social sciences, especially those affected by analytical philosophy, are less vulnerable to this mistake, since they tend to carefully scrutinize their definitions and claims.)
These problems have to do with the subject matter, however. It is normal that true problems within the social sciences are hard to come by: there is little that can be said with the certainty or precision of the sciences. At a sufficiently general level, there are plenty of falsifiable (and therefore potentially true) claims to be made. Take US society: the US war in Vietnam killed over a million Vietnamese civilians; the US is a freer country than China; there are more obese people today than a century ago. So far, no problem. But when the social sciences/humanities begin making claims about human affairs in highly specific terms, that is where they often err.
One learns of the Christian or Muslim mind; the comparative advantages of capitalism; the nature of dreams; the violence within language; the aesthetics of fascism; the instability of logocentric thinking; the path of History; the nature of queer subjectivity; the limits of the episteme; the politics of anthropological thought.
Once one has divorced oneself from having to actually provide evidence to support a precise claim, practically anything can be claimed, free of conceptual coherency and relation to the world of establishable knowledge. The claim might mean something, or not (Western metaphysics is phallogocentric; rhizomatic knowledge structures are democratic; the unconscious is structured like a language; our regime of knowledge is contingent; violence inheres in language; dreams are wish fulfillments; art is wish fulfillment; art expresses society’s inner nature; desire is mimetic).
The main point I want to make here is two-fold: first, that there are fewer real problems to be found in the social sciences, so people are constantly making them up.
Second, that because of the difficulty in locating coherent, stable, rationally observable problems, there often is little to unify a field–and therefore little knowledge to accumulate. Rather than continue studying a problem, researchers merely reproduce trivial observations upon a new terrain. So this is the data-collection stage in science, the most superficial. But the social sciences never get beyond it, because they often don’t have a stable object or a problem to solve. For example, you just pick a different novel to read, and apply some newly invented concept: the rhizomatic structure of Irish fairy tales; the spectatorship of desire in the 18th century French novel. The field is thus reduced to endless data collection, without ever coming to say something, to solve or even define a single problem.
Again, this is for a good reason: there aren’t many problem to solve here; there is very little solid knowledge to be had. This doesn’t mean these phenomena (human affairs) aren’t worth studying–they are; it just means that claims being made should be much more cautious and humble; the researches much more careful in their methods; and the problems defined in a clearer way.
One way to test what I am claiming here is to see if the problems have shifted since, say, Aristotle; and if we are any closer to developing knowledge about them. Take dreams. It turns out that Aristotle’s claims about dreams are about as valid as Freud’s. Centuries passed, knowledge did not accumulate, and the problem is pretty much the same one, in the same state it was 2,000 years ago.
Take novels, plays, poems, and literature. A lot more of them have been written, and about plenty of new things since Aristotle. But our understanding of literature is not more advanced since the Greeks; Aristotle’s Poetics is as insightful and advanced as anything from a book today on “Theory of the Novel” or “Poetics.”
Take democracy. Here there has been a bit more nuance added to the political ideas that were around in classical Greek society, but very little. Any book today on democracy usually starts and ends with the same concepts developed in Greek and Roman society. After two thousand years, our ideas about politics are pretty much the same ones.
One could go on like this, but it’s not necessary. Compare, on the other hand, to the sciences: observable phenomena of nature. Here, great leaps have been made; explanatory and predictive theories have been developed, tested, proven; major advances in technical ability have been accomplished.
Prophets, cults, and trendy ideas
Since the social sciences can’t usually produce reliable, testable, provable results, they often substitute trendiness, fashion, and prestige. What one finds is that every generation or so produces its set of “most fashionable ideas,” which is then dumped for the next set after a decade or more. And this is without getting any closer to developing any coherent explanation for a coherent problem.
There is a mode of cultishness in this. Suddenly, to be “in,” one finds that one has to use a set of pre-approved terms and concepts. Not because these ideas are of special explanatory value, not because they contain special insights, but because they happen to be the current trends. Some of the recent terms that it is fashionable to use include: homo sacer, panopticism, governmentality, jouissance, postcolonial, subaltern, ocularity, rhizomatic, deconstructive. When new, approved terms become available, they are snapped up by eager grad students and young faculty seeking to gain the prestige and awe produced by using the newest terminology. Note that we are far away from realizing insights or improving the state of knowledge; at this point we’re just playing a fashion game.
Reasons: institutional incentives
I want to point out three structural reasons for this.
1. Arbitrary Department Divisions
First, the cloistering of departments in the social sciences around ridiculously arbitrary distinctions. Literature studies are supposed to be their own kind of knowledge, distinct from history departments; which are distinct from sociology departments; economic departments; anthropology departments; and so on.
In fact, what are distinct are the problems pursued, not the kinds of knowledge. Rational inquiry operates the same way everywhere: you define; you look for evidence; you present arguments.
Departments ought to be organized in a much more collaborative fashion; this would immediately address some of the bubbles of approval that get created on some campuses, where common sense in one department is a total mystery in the next. It would also produce a stimulating mingling of people with different orientations, who at their best could help each other avoid the common pitfalls. Historians talking with literature scholars; sociologists with political scientists. They would be oriented around common problems, just like the sciences; and not around arbitrary divisions. So, everybody working on Algeria would be in contact and would be able to work cooperatively; everyone working on the history of political arrangements could work together (note that this would be relevant to historians, literary scholars, anthropologists, sociologists, political scientists, etc.).
2. Publish or Perish
This is a real recipe for disaster in the social sciences. In the sciences, publishing is just a way of updating people on your research. It’s not about trying to make claims beyond what you can prove; it’s about putting the results of some experiment out there so other people can use them, compare your work to the next guys, and decide how to improve the experiment or modify the explanation.
In the social sciences, it’s nothing of the sort. There is no “lab head” whose established, recognized wisdom is available to organize the lab’s next experiment in a manner coherent with the state of the field, the probability of getting results. And there is no coherent set of problems the field is working on or working out or stumbling over. Just the same old concepts as always: politics, culture. So since you have to publish something to prove yourself, and there’s no real rhyme or reason to it, you just pick something that sounds fun, interesting, or impressive. And your ultimate goal is not to present knowledge, but to impress.
Writing to impress is clearly at conflict with writing to present honest results. Suddenly there are incentives to make the most grandiose claims possible, use the most prestigious discourse, make your argument seem impressive, instead of accessible.
3. Institutional hierarchy
The hierarchy present in the social science departments are virtually going to guarantee a reproduction of prestige instead of original, honest, creative thinking. People are not free to disagree with their superiors, at least not outside of a pretty strictly-defined parameter, because, given the state of the social sciences described above, there are no rational standards by which to prove one claim right and the other wrong.
The result is that in order to get promoted, you are strongly incentivized to tell your superiors what you think they want to hear; to tell your peers what you think will impress them the most. This is a far cry from being motivated to say things the way you see them; disagree with anything that doesn’t make sense to you; explore your own ideas even if they are far from the fold.
This model is not unlike what Galileo was doing: when he said the wrong thing, he got ejected from the system, summoned to the Inquisition the way a social scientist gets summoned to the Chair’s office and told he needs to get in line or get ejected.
This is the opposite of a healthy environment for coming up with good ideas. The hierarchical nature of these institutions strongly incentivizes group-think, emulation, and approval-seeking. A major advance would be made if people, once they had met the conditions required to be considered a professional quality researcher, were no longer under requirements to produce knowledge content that is approved of, but instead could simply say what they think.
Here are the article titles of the last two issues of Diacritics:
–The Society of the Spectral
–The Turn Away from Marxism, or Why We Read the Way We Read
–The Brechtian Exception: From Weimar to the Cold War
–In the Theatre of Politics: Althusser’s Aleatory Materialism and Aesthetics
–Huntology: Ontological Pursuits and Still Lives
–Acts of Askesis, Scenes of Poiesis: The Dramatic Phenomenology of Another Violence in a Muslim Painter-Poet
–What a Wall Wants, or How Graffiti Thinks: Nomad Grammatology in the French Banlieue
–Image and Silence
 As an example of American culture studies, take the following journals as a representative sample: Critical Inquiry; Diacritics; Modern Language Association.