July 06, 2004

Comments on comments, again

[Note, 3/15/2005: as a result of an appeal from the person who wrote the first comment discussed below, I've anonymized the names of the commenters involved. ]

I've just removed two comments from an earlier post. They haven't been censored -- you'll get a chance to read them, if you want -- but they're no longer where they were. The rest of this post explains why I moved them. I hope that the explanation will make sense to you, and will help make interventions of this kind less frequent in the future.

Yesterday, Geoff Nunberg posted a critical evaluation of the recent Groseclose and Milyo paper on measuring political bias in the media. The G&M paper has been widely discussed among social scientists, as well as in the popular press and the blogosphere. Geoff's essay explains several important problems with the G&M study that I haven't seen presented elsewhere, and it does so in a way that will be accessible to a wide audience. I'm proud to have it in our blog, and I expect it to be widely read and discussed.

Shortly after Geoff's post went up, Xxxx Xxxxx started off the comments with a 600-word essay that began "Let me shift the issue a little bit. In Generative Linguistics done in Brazil there is one ideological bias that is undeniable: the segregationist bias. This idea is not something Brazilians came with. It started with foreigners that began to study 'Brazilian' Portuguese". Xxxx went on to present a long discussion of a controversy in Portuguese dialect studies. You can read the whole thing here. Whatever the thought process that connects this stuff to Geoff's critique of the methodology and implementation of the G&M media bias study, nearly all of it is on a different topic, one that most of the readers of Geoff's essay will find puzzling at best.

Shortly after Xxxx's comment went up, Yyyyy Yyyyyyy responded with a comment of his own, which is reproduced in the same file. Yyyyy's comment was very much on topic -- that is, on Xxxx's topic -- but there was by now no glimmer of a connection to Geoff's topic. Again, a reader who had come for a serious discussion of the G&M media bias paper would be baffled by all this .

The effect of these two long comments would certainly be to prevent most other readers from commenting on the actual content of Geoff's post. Worse, this stuff may give readers the mistaken impression that Geoff (is a member of a community that) believes that the G&M media bias paper is somehow about the ideology of Portuguese dialectology.

So I've removed those two comments.

We welcome comments -- well, to be frank, there's some difference of emphasis if not opinion on this point among us -- but at least, some of us welcome comments. You'll help us to keep the welcome mat out if you can stick to the topic, more or less, and be reasonably brief. If one of our posts reminds you of something interesting but completely different, please blog it yourself and tell us about it, and we'll probably post a link. If you have a lot to say about one of our posts, please put your essay up on your own blog, or your own website, and put a brief description and a link in the comments on ours.

I recognize that topic drift in comments is sometimes a creative process. But this was a carefully researched and focused post, on a topic of broad interest, and the very first comment hijacked the discussion and took it off in a completely different direction, at great length. That's just plain rude.

Posted by Mark Liberman at July 6, 2004 02:05 PM

Sorry. Attitudes towards that sort of thing vary from blog to blog. I won't do it here again.

Posted by: Scott Martens at July 6, 2004 04:38 PM

Sometimes an example or a comparison is long to show a point.
Again, it has already been observed that showing something is stronger than just saying it. And the idea of my example was to show how an ideological bias works in a certain 'medium'. It was not my intention to direct the issue to another discussion. I could have given another example like the anti-Chomsky bias of some European circles.

Posted by: Tony Marmo at July 6, 2004 05:21 PM

Hi Scott,

> Attitudes towards that sort of thing vary from blog to blog.

True enough, and we haven't really developed a comments culture here yet.

Posted by: Mark Liberman at July 6, 2004 07:34 PM

Hi Tony,

>... the idea of my example was to show how an ide logical bias works in a certain 'medium'.

The trouble is, the topic of Geoff's post wasn't ideological bias, whatever the medium. It was the methodology and execution of the Groseclose and Milyo study, and the implications for its validity.

> I could have given another example like the anti-Chomsky bias of some European circles.

That would have been just as off topic.

Posted by: Mark Liberman at July 6, 2004 07:40 PM

But the Groseclose and Milyo study does not seem to aim nothing but to denounce an ideological bias they claim to exist. Reading their study with critical attention, you will notice that they care less about methodology or execution.

The large mass media in the US are much more sympathetic to Republican candidates and republican ideas than to Democrats and their ideas. The thing is that, as one German marketing and propaganda expert once explained, the best campaign is the one that does not look like a campaign, it is the commercial that does not look like a commercial. So if the republicanism of the mass media is to have its intended effects, the republicans have to pretend to believe that the media tend to diffuse liberal ideas, when it is rather the contrary.

Now, to understand how this works one has to look at parallel examples in depth, which requires detailed description of such parallels. Examples are not offtopic, they are just examples and in order to understand one text the reader must keep in mind this. But if you take an example to be the starting point of an entirely different discussion, then you fail to grasp the general idea.

Posted by: Tony Marmo at July 6, 2004 08:59 PM

Tony, if you want to "look at parallel examples in depth", please feel free to do so on your own interesting blog
Other bloggers, including us, will take account of your contributions as part of the discussion.

In the comments here, please try to say something relevant to the post's topic. In the case of Geoff's post on the Groseclose and Milyo paper, his topic was their methodology and its execution, not the nature of bias or of campaigns against it.

Off topic comments, especially long ones, are liable to be deleted.

Posted by: Mark Liberman at July 6, 2004 09:18 PM

But what I said is more than relevant to the post's topic.
How do you discuss a topic in linguistics? If someone provides you data and analysis, don't you respond with more data and analysis. So if you give me one sentence and analyse it, and I give you back another sentence and another analysis, is it offtopic?

By the way, I do not think my blog is interesting. If my blog was so interesting it would already be among the ones you list. Haha!

Posted by: Tony Marmo at July 6, 2004 09:22 PM

PS: I had reposted the final part of my comment not mentioning anything about Brazilian Portuguese, which made my comments much briefer and yet it got deleted in the same way.

Posted by: Tony Marmo at July 6, 2004 09:26 PM

Tony, your weblog is in our blogroll on the right, and if you post something relevant on media bias, we'll link to it specifically, as no doubt will others.

I don't have anything more to say on this, and additional off-topic comments will also be deleted.

Posted by: Mark Liberman at July 6, 2004 09:40 PM

Thank you very much.

But why a concise reply about 'conservative versus liberal' bias to a post which, as indicated in its title, is about the presumed 'liberal bias' offtopic?

Posted by: Tony Marmo at July 6, 2004 10:02 PM

Please, an explanation rather than an arbitrary act of deletion. Geoff Nunberg has a right to recieve posts from people that agree with him, and dissent from openly conservative inquisitors.

Posted by: Tony Marmo at July 6, 2004 10:14 PM