色盒直播

同行评议能否在社会科学的范式战争中幸存?

如果作者不再需要证明他们的基本假设是正确的,那么裁判员该怎么办?马丁·哈默斯利(Martyn Hammersley)问道

七月 19, 2022
Footballers arguing with the referee in the centre to illustrate
Source: Getty (edited)

点击阅读英文原文


同行评审对研究界的运作至关重要。但是,它必须满足哪些条件才能运行呢?其中一项答案是肯定的,对于研究任务以及如何追求它,必须有一些最低程度的一致意见。

然而,在社会科学和人文科学的许多领域,不仅对所研究内容的性质以及如何理解它,甚至在研究的产物应该是什么方面,也存在着越来越多的根本分歧。例如:人类、组织和机构是以因果关系在世界上运作,还是以在“文本之外”没有任何顺序的散漫结构运行?研究的目的是了解世界还是对世界产生“影响”,包括减少社会不平等?

这里涉及的是基本承诺之间的冲突,导致了“范式战争”,内特·盖奇(Nate Gage)在1989年就了其未来。同行评审是否与这些冲突兼容?让我来说明一下这个问题。

最近,在为一家期刊审阅一篇论文时,我遇到了一个两难境地。这篇论文提出了一些有趣的观点,但我的观点是,它依赖于一系列可疑的经验、理论和政治假设,导致对相当薄的数据进行有倾向性的解释。作者暗示,对这些假设的任何质疑都相当于对他们智力和社会身份的攻击。但我觉得,既然很多读者不会分享它们,就应该提供明确的理由。

鉴于此,我建议进行重大的修订。其他审稿人则更为积极,尽管他们确实发现了各种需要注意的问题。编辑们决定进行小的修改。重新提交的文件没有对我的评论作出任何实质性修改。更重要的是,他们的附信根本没有回复我,只回复了另外两位审稿人。

编辑们随后要求作者回答我的评论,但他们的回答是,由于他们有不同的“认识论立场”,他们不需要对我提出的具体观点提出反驳。他们声称我们之间的差异是“不可调和的”,超出了他们论文的范围,我的“立场”影响了我“对'有效数据'的个人看法”。他们坚持认为,他们的“方法”已经过“学院彻底审查,是可靠、有效和值得信赖的”。换句话说,由于他们能够诉诸于分享他们承诺的文献,因此没有必要在论文中证明这些理由。

我不同意并建议予以拒绝,特别是因为在我看来,这篇论文仍然存在根本性的缺陷。然而,在第二阶段,另一位审稿人推荐出版该论文并评论道:“感谢这些作者对一号审稿人的抵制。 ”

“反击”听起来像是一个军事隐喻;言下之意似乎是,仅仅参与我所提出的关键点,就等于在面对需要击退的东西时投降。请注意,这里争论的不是作者没有修改他们的论文;同行评审不需要这样做。问题在于,他们起初拒绝做出任何回应,当被追问时,他们只是诉诸于他们自己的范式承诺。它奏效了。这篇文章即将进行发表。

当然,可以说,我应该承认作者声称遵守的替代范式的合法性,并接受我的批评性评论因此不适用,只是反映了我自己的范式。许多人会争辩说,欢迎具有多样性的方向是学术生活的一项原则,我在某种程度上同意这一点。但宽容显然是有限制的——事实上,这里讨论的争端的双方都在实践中认识到了这一点。问题是:应该根据什么理由确定这些限制?

作者实际上拒绝了支持同行评审的关键假设。他们采用了后来被称为立场认识论的一个版本,根据该版本,一些观点被赋予可信度(而另一些观点则被剥夺了可信度),基于提出它们的人所属的社会类别。边缘化或受压迫群体的成员(或者更确切地说,那些声称代表他们发言的人)的观点被赋予了认识特权,而那些被认为是主导群体的成员所持有的观点则被拒绝,例如(充其量算得上是)“意见”。

因此,一方面,可接受的限度是根据作者自由承认致力于政治目标的范式来定义的;而就我而言,限制来自学术同行评审所需的内容。这种冲突源于这样一个事实,即同行评审要求所有“同行”都被视为平等,而不是认识论上的特权。

事实上,匿名被用来使作者和审稿人尽可能地对彼此的社会特征和政治观点视而不见(除其他事项外)。重要的是他们是同一研究界的其他成员;他们必须仅在此基础上相互接触,而不是根据政治或范式承诺。

在同行评审中放弃这一要求可以容忍吗?当然,它从来都不是完美的,但如果同行评审只是范式战争的另一种工具,那么它仍然是合理的吗?

马丁·哈默斯利(Martyn Hammersley)是开放大学教育和社会研究的名誉教授。

本文由张万琪为泰晤士高等教育翻译。

后记

Print headline:?Social science’s paradigm wars imperil peer review

请先注册再继续

为何要注册?

  • 注册是免费的,而且十分便捷
  • 注册成功后,您每月可免费阅读3篇文章
  • 订阅我们的邮件
注册
Please 登录 or 注册 to read this article.

Reader's comments (10)

Very interesting article, thank you Martyn. I'd suggest this is further complicated by the more old-fashioned* politics of, for example, supporting and promoting your faction against rivals; crushing and marginalising personal rivals; etc.. * Only kidding, I know they're still alive and fully at play here.
Good grief. I can only thank the good Lord that I'm not a social scientist. Long live Sokal and Bricmont
1) The Sokal affair paper was in a non-peer-reviewed journal. 2) Sokal and Bricmont‘s book only sets out to address a) the perceived misuse by a select set of authors of terminology from mathematics and the natural sciences, and b) a very particular subset of critical sociology of science which they saw as going too far to be useful. Both, particularly the latter, are important and worth discussing, but neither are fundamental flaws which somehow doom the social sciences and humanities.
The Sokal paper was indeed published in Social Text, which describes itself as governed by an editorial collective. Presumably this means that more than one person read the hoax paper. Peer review by another name?
Is it common to have editors of a journal simply shrug off a reviewer's comments like that and accept that the writers were immune from criticism because they had a privileged 'onto-epistemological position'? Is the whole journal now committed to that same position? As someone once said -- I think we should be told.
In a world where tweets and tabloid headlines seem to influence public policy and opinion more than serious social science, and where arguably most academic publishing has little or no impact, other than perhaps the career trajectories of the authors, I wonder if there is a deeper problem than simply peer review. The case for standpoint-free, objective social science as the way to counter fake news, seems like a sensible idea. Yet, history would suggest that many of the established theories of social science are tainted with all kinds of ideological baggage. So, until and unless we can devise a cast iron basis for objectively determining the quality of academic publications, I think accepting and making one's standpoint offers the reader a clearer way to evaluate the work being reviewed. Hence, rather than dismissing standpoint theory, shouldn't all research be assumed to be from a particular standpoint which needs to be acknowledged?
Gurnam Singh is right, of course, but standpoints vary according to their openness and the strength with which they are held. Martyn Hammersley is defending a standpoint based on academic argument and notions of evidence which makes his views open to correction in principle at least. A suitable response would have met his objections. The authors of the article seem to be adopting a standpoint that is rooted in their personal identities and political commitments that is far less open to argument and correction: any attempt to question their conclusions must only be an unwarranted personal attack aimed at their strongly-held beliefs, probably even personally or politically motivated. The test for me would be -- is there anything at all that the authors would accept that would invalidate their conclusions?
Couldn't agree more with Darris. I may feel the earth is flat but it is not true. There are many areas of social sciences, arts and humanities that make a valuable contribution to our existence often through invaluable critique of the paths down which we tread. The issue, however, seems to boil down to what counts as research or excellent scholarship and what should be labelled personal opinion. Everybody is entitled to a personal opinion and indeed it is valuable to chronicle these in diaries and or perhaps even a Journal of Personal Opinion. Of themselves, such writings provide useful data as a reflection of the times and as new data for future scholars. The extent to which we should pay any attention to such opinions now should be determined by whether or not a set of clearly identified and agreed criteria have been applied to the collection of information, its synthesis and process of interpretation. One such basic criterion could be applied as a minimum standard is whether or not the author is able to satisfy a standard of reliability - in other words given the same data (however you define it - and here are rues around that) would the overwhelming majority of humanity come to the same interpretation and subsequent conclusion. Upon such things societies grow and exist relatively harmoniously. This doesn't mean that the majority are always correct in their attributions of causality but, using an accepted methodology, what we used to think was true can be refined accordingly when new evidence rolls in. Without demonstrated reliability in the methods there can only be (at best) minimal validity. If the opinions are focused solely through a lens of identity and/or political ideology they are likely to be characterised as being unreliable and therefore of dubious validity - but how can we be sure? Kahneman and Tversky's work on the biases and heuristics of the human information processing system clearly show how fallible we are as processors of information. Left unchecked by a clearly articulated methodology, designed to arrest our fallibility as information processors, our thought processes can only lead us in many (fruitless) directions like the preverbal headless chicken. The value in research is it should seek to answer a question and there are accepted ways of doing this. If you don't start with a clearly articulated question how can you do research? Without a clearly articulated question the activity is clearly something else and should be identified for clarity, honesty and integrity (it's ok ato ask what should the question be - this too can have an acceptable methodology). Some authors may believe passionately what they are writing is true (indeed some openly identify as activists) but in the court of informed public scrutiny it may not stack up in the collectively agreed reality and appears more as a cult or a new form of minorityreligion. Unchecked by openly articulated methods, we soon end up heading down various rabbit holes as some corners of academia demonstrate on a daily basis. At issue, surely, is how as a society, do we create new usable and meaningful knowledge some of which some might have direct application or some may used to generate new questions hopefully for a purpose with an end point in mind. Making a series of un-falsifiable statements based on untested assumptions is really not helpful nor should tax payers have to pay for it, or indeed the self-serving echo chamber from which it comes. Other areas, sadly many of which would be described by the many hues of post modernism, provide opinion dressed up, sadly, in the overcoat of respectability afforded by faculty membership. If it doesn't add value, which could be referenced in myriad ways, to those who fund it, why should it continue to be supported? Politicians and vice-chancellors, in particular, should be asked how the public is getting a valuable return on its research investment with respect to what often masquerades as research when it clearly is not.
I am very grateful to all those who have commented on my article. The discussion raises important and difficult issues that need addressing. I am especially indebted to darris for the accurate and concise summary of my position.
Did Donald Trump steal documents from the US government? That should be an answerable question as long as you have an agreed and clear definition of "steal". It should not matter on how you feel about the matter. If you do not have an agreed concept of truth, any bully can have their way and trample on their fellow citizens. At the moment, Judge Eileen Cannon is a threat not only to US security, but to all of us throughout the world. A world or unreason is a world of ignorance, pestilence, and worse.
ADVERTISEMENT