Re: Why do you believe what you believe?
- From: Burkhard <b.schafer@xxxxxxxx>
- Date: Wed, 30 Sep 2009 14:44:05 -0700 (PDT)
On Sep 30, 10:28 pm, Garamond Lethe <cartographi...@xxxxxxxxxxxxxx>
wrote:
On 2009-09-30, Burkhard <b.scha...@xxxxxxxx> wrote:
On Sep 30, 8:02 am, Pulse <feralpu...@xxxxxxxxx> wrote:
<snip>
What makes specific secular texts more reliable than other contradictory
texts? Here I'm mostly interested in understanding why many consider
peer review journals so valuable as well as which (if any) other texts
can be considered of equal merit.
OK, here my part 2:
<snip good stuff>
....and very good it is, but I think there's a bit too much bias here towards
empirical papers. Lamport's logical clocks paper does fit this mold at all.
I just taught Culler's LogP paper last week, and at this distance it's obvious
that the empirical results were just window dressing to make the model look
good.
I think you've done a great job of capturing the easier (and smaller) aspect
of the problem: reproducibility. Good results are (at least in theory)
reproducible, and we have many proxies to choose from because reproducing
experiments is expensive (in term of cost, time and career advancement).
The much more interesting problem (to me at least) is how does one judge
quality on issues where honest, knowledgeable people disagree? For
example:
1. Do we need a model to explain this (is the problem important)?
2. Does the proposed model allow us to answer interesting questions?
3. When is it worthwhile to publish a more complex model that makes
slightly better predictions?
4. When is it worthwhile to publish a model that makes better predictions
for a smaller set of parameters?
The only answers I've come across have been from reading lots of papers,
reading reviews of my papers, and reviewing other people's papers under the
guidance of my adviser. To a degree that might upset non-scientists here,
I think it is mostly a matter of taste. I need to know the tastes of the
community for whom I am writing, as my reviews will help influence that
taste.
In short, it's messy.
Oh, absolutely. The good thing, if you can call it like this, is that
the evidence shows that in the non-empirical disciplines, almost all
papers get published eventually. E.g. I recommended a paper for
rejection last year, not because there were any mistakes in it, but
because in my opinion it did not say anything that had not been said
better before. I just read it, published, today - the author has
downgraded journals, and actually taken some of my comments on board.
My favourite example of a rejected (conference) paper was a submission
which sounded eerily familiar - it was a copy of a paper a friend of
mine had written 15 years ago. (A visa scam, with other words, they
just needed a letter er of invitation)
It is indeed very rare that I find actual mistakes. Omission of
relevant prior research is in my experience the most obvious problem,
followed by exaggerated claims about usefulness/impact. (we solved a
minor problem in proof theory - this means tomorrow we will have full
blown AI systems as a result and maybe also solve global poverty ...)
But the thing is, in these fields the role is indeed less a gatekeeper
of science, but a facilitator of science communication -making sure
the article ends up in the appropriate channel. Which his why I often
recommend other journals when I propose a rejection.
.
- References:
- Re: Why do you believe what you believe?
- From: Garamond Lethe
- Re: Why do you believe what you believe?
- Prev by Date: Re: The TO'sFIRST evolutionist of the month vote
- Next by Date: Re: Resident loonies
- Previous by thread: Re: Why do you believe what you believe?
- Next by thread: Re: Why do you believe what you believe?
- Index(es):
Relevant Pages
|
Loading
