> According to a survey of the views of members, 75% described themselves
> as non-religious. Just bringing up religion on a transhumanist forum
> tends to start fierce denunciations of it as an evil, death-bringing
> meme. On the other hand, transhumanism certainly deals with issues
> traditional religion and spirituality are interested in: the
> perfectibility of humans, immortality and eschatology. Who can deny
> there is a certain millennialism in some of the writings about the
> Singularity?

Who can deny there is a certain gibberish in some of the writings about quantum mechanics or Godel's Theorem? I am going to keep on hammering on this point: There are people who only wish to relax with some comfortable gibberish, and what such people say about a theory does not impugn the integrity of the theory. The bad writings are simply ignorable, eliminable from the domain of arguments, and only if there is no rigorous writing remaining does the theory have a problem.

Ignoring bad writing presupposes that it is easily identified. Fields in trouble can have plenty of useful writings, but they cannot be discerned in the mess. See
http://www.aleph.se/andart/archives/2004/01/scholarship_and_its_antibodies.html

Many of the academic codes have developed as a first filter; if you don't cite the literature or use the right terminology you are likely not familiar with the field and hence unlikely to have anything relevant to contribute. But if further scanning cannot easily tell you whether the results or arguments hold water, then the effort to weed out bad stuff becomes too expensive. Which makes outsiders, trying to judge the field, unwilling to wade in to see if there are a lot of gold nuggets there, or just dross.

If one wants to ensure that a result one believes in gets noted, one better make it's quality notable through some hard-to-fake signal.

It is really so difficult to identify bad writing? It seems to me we can still come to reasoned conclusions in a number of difficult cases; granted, the effort needed makes peer review clunkier, but well worth it (indeed, necessary).

Mathematicians, for example, seem to have understood this problem earlier than other specialists. The Clay Institute, with its million dollar prizes for the solution of major problems, includes as part of the assessment rules that not only must your solution to one of the problems be published in a respectable math journal, it must withstand at least two years worth of scrutiny from the mathematics community. Only after at least 2 years, and given that the math community is in broad agreement that you've solved the problem, will you be considered successful. (If 2+ years go by, but there is a plausible objection or non-negligible controversy, then you'll have to wait until the ambiguity is worked out.)
Similar widespread and lengthy scrutiny was applied to Wiles' proof of Fermat's Last Theorem. Even now that the details are all worked out, when mathematicians who understand the proof were asked about the chance of laypersons understanding it (esp. with respect to the chances of Fermat himself having used Wiles' sort of proof), they responded that it was impossible because the proof (over a hundred pages long) required *even other mathematics research professors* to devote months of study before they could understand it.

The evaluative dilemma only appears to be a problem with highly complex subjects. Hence the Clay Institute's policy, hence the confusion and debate over the Bogdanoff brothers' novel application of topological field theory to Planck scale Big Bang cosmology, hence the Peter Lynd controversy.

Singularity theory, I dare say, is not quite on par complexity-wise with the boundary-pushing math & physics examples above (not yet anyway), which is why it is much easier by comparison to filter out the good writing from the bad.

Coming upon cases where bad writing is hard to identify does not indicate bad writing is always (or even most of the time, or half the time, etc.) hard to identify, nor does it indicate bad writing is equally hard to identify in each field. Thus, I would say Eliezer's suggestion to ignore bad writing should presuppose bad writing to be identifiable within reasonable limits. To do otherwise would be to presuppose in the other direction, and with less justification, as we have many more examples of easily-identified bad writing than we do of hard-to-identify bad writing.

I agree with the motivating background concept here and in "scholarship & its antibodies," though; the problems of scholarly evaluation definitely warrant much greater study than they have received. As the complexity of subject matter continues to increase, so too will the necessity of finding better alternatives that enable growth of scholarship.