Machine translation press coverage or no-news evangelism

von Peter Winslow, veröffentlicht am 15.09.2022

A recent Slator piece titled “Can Machine Translation Facilitate Better Global Meetings in a Common Language?,” published today, September 15, 2022, seems like a good opportunity to address some general features of the press coverage that machine translation gets. Like other things I’ve discussed here in the past, this coverage strikes me as paradoxical.

That piece starts with these words:

A group of researchers set out to prove that team members who are globally distributed can better prepare for general, common-language meetings by sharing information ahead of time.

This sentence is extraordinary. Paradoxically, it invites readers to believe a statement that cannot possibly be true. Anyway, that statement cannot possibly be true absent extraordinary evidence. And it is unclear what that evidence could look like: has any group of researchers ever doubted that sharing information ahead of time can better prepare team members for general, common-language meetings, be those members globally distributed or not? Has anybody ever doubted that? If this sentence were true, the real story would not be about machine translation and whether it facilitates global meetings or no; the real story would be about the truth of that sentence and what that truth entails.

I know this is a somewhat unfair representation of Slator’s article, or whatever one wishes to call it. It is also a somewhat fair representation still. That article doesn’t get much better (among other things, it seems to confuse structure with clarity, but that’s a story for another day). Eventually one does learn that by “globally distributed” the author means little more than “bilingual” or “multilingual.” I suppose, then, a charitable reading is that “globally distributed” is an instance of the rhetorical figure metonymy. Words matter, beginnings matter, rhetorical figures matter, and machine translation coverage would benefit greatly from more sober, and sobering, prose. It demands too much from its readers. Indeed, on its face, it strains credulity too often.

It strains credulity too often, because it is too often very good – I say often, because it is occasionally bad. But I do not wish to give evidence of what “good” means here, what “bad;” neither do I wish to address the specifics of this coverage. Rather, I wish to address some general features of that coverage. And a general way to think about that coverage is as the paradox that it is, which is to say: as a kind of “no-news evangelism.”

Put differently, the good press coverage of machine translation resembles a certain kind of bad evangelism.

  • Paradoxically, good press coverage of machine translation often fails to deliver any good news; indeed, it often fails to deliver any news at all.
  • Where it does deliver news, it often invites readers to confuse the excitement surrounding that news with the news itself.
  • The excitement renders difficult, if not impossible, any clear assessment of the purported news.
  • I do not mean to suggest the message is distorted; I mean the excitement is the message – it’s fallacious, an appeal to emotion.
  • Where this coverage does not appeal to emotion, it is not exactly dishonest, but not exactly honest either: it often presents marketing shticks as expert – anyway, as trustworthy – opinion.
  • There often exists a tension between positive and negative statements about machine translation; the positive statements are positive in the extreme (“MT Logs Added Clarity, Depth, Comfort”), the negative statements are almost not negative at all; rather, they are almost always unduly tempered (“MT and Other Interventions Render Mixed Results”).
  • In other words, nothing ever really counts against machine translation; press coverage of machine translation always has at least something positive to say about machine translation – anyway, it always mentions something that counts for it.
Diesen Beitrag per E-Mail weiterempfehlenDruckversion

Hinweise zur bestehenden Moderationspraxis
Kommentar schreiben

1 Kommentar

Kommentare als Feed abonnieren

I am aware that this portrayal of Slator's piece, or whatever you want to call it, is quite unjust. In some ways, it is still a fair representation. It doesn't get much better after that (among other problems, it seems to conflate clarity with organisation, but that's a topic for another day). One does eventually discover that the author is referring to "bilingual" or "multilingual" when they say "globally diffused." The phrase "globally spread" could, I think, be interpreted in a charitable manner as a rhetorical figure metonymy. More serious, sobering literature would considerably improve machine translation coverage because words matter, beginnings matter, rhetorical figures matter, and so do they for machine translation. It asks too much of its audience. In fact, on first glance, it frequently strains belief.

Kommentar hinzufügen