The trouble with making linguistic claims

« previous post | next post »

There is a lot for reasonable people to agree with and disagree with in Philip Kitcher's recent essay in The New Republic, "The Trouble with Scientism: Why history and the humanities are also a form of knowledge". This being Language Log, however, I can only urge readers of Kitcher's essay to take the following linguistic claim with a healthy dose of skepticism:

In English we speak about science in the singular, but both French and German wisely retain the plural.

Kitcher's point in making this claim — and the actual, reasonable argument that follows it — is that "science" is hardly a singular thing:

The enterprises that we [English speakers–EB] lump together [with the singular word "science"–EB] are remarkably various in their methods, and also in the extent of their successes. The achievements of molecular engineering or of measurements derived from quantum theory do not hold across all of biology, or chemistry, or even physics.

This argument is a key part of the larger (and again, reasonable) argument laid bare in the essay's subtitle: that "history and the humanities are also a form of knowledge". Anyone interested in this kind of topic (as I am) is encouraged to read this essay, followed by the other links further above, and perhaps counterbalanced by this NYT Opinionator blog post. (And don't forget to squeeze the comments.)

So what about the linguistic claim? Unfortunately for Kitcher, it's complete hogwash.

Well, OK, let me soften that: as stated, at least, it seems to me to be complete hogwash. As a native English speaker who "speaks about science" with some regularity, I can attest that I also speak about "the sciences" with some regularity. And I'm not a native speaker of French or German, but I have passing knowledge of both (four years of grade school French and two years of college German); I'm fairly certain that French speakers speak about "la science" (feminine singular) as well as "les sciences" (plural) and that German speakers speak about "die Wissenschaft" (feminine singular) as well as "die Wissenschaften" (plural).

So what is Kitcher on about? It may be the case (and I stress that it is far from obvious) that the plural forms predominate in French-language and German-language academic discourses on the topic, and that the singular form predominates in English-language academic discourse — but why would anyone want to elevate such an (entirely hypothetical) asymmetry to the level of facts about the languages themselves? [ Update: see the discussion in this comment below. — EB ]

I'll end this post with a bit of a cheap shot: searching the text of Kitcher's own (English-language) essay, there are 37 instances of the string "science", 20 (more than half!) of which are embedded in the larger string "sciences". Now to be fair, Kitcher is specifically interested in talking about "sciences" as opposed to "science", but I honestly don't think any English speakers would bat an eye at his "overuse" of "sciences". Kitcher could easily have made his point without making a factually incorrect linguistic claim.


  1. Aaron said,

    May 30, 2012 @ 11:03 pm

    Surely he wouldn't claim that the difference between the American "math" versus the British "maths" carries any deep meaning? Sometimes a plural is just a plural.

  2. Gabriele Zöttl said,

    May 30, 2012 @ 11:15 pm

    Hi Eric,

    just to confirm your doubts about the usage of German 'Wissenschaft/en'. We (not exclusively, but mostly) use 'Wissenschaft' in the singular form when we are talking about science as such or the whole scientific community.

    There is, however, a trickier semantic point which explains our plural, invalidates Kitcher's German argument and makes his entire essay somewhat hard to convey to a German audience: English 'science' is restricted to natural sciences, while German 'Wissenschaften' include everything that is explored with scientific methods … literature and linguistics, history and sociology, you name it. So German readers will have a hard time understanding why anyone feels a need to explain why 'history is also a form of knowledge'.

  3. Tom O'Neill said,

    May 30, 2012 @ 11:23 pm

    But consider this: Sciences Po was established in February 1872 as the École Libre des Sciences Politiques by a group of French intellectuals, politicians and businessmen. I don't believe a single university in the USA has a Department of Political Sciences.

    [ This last point may well be true, but a quick search for {"department of political sciences"} reveals lots of ad hoc references to departments of political sciences in the USA (and many more elsewhere), even though the relevant department names seem to all be in the singular. And I'll add that these departments tend to be housed within divisions/schools of social sciences or arts and sciences, both plural. — EB ]

  4. Kevin said,

    May 31, 2012 @ 12:51 am

    Although Departments of Social Sciences seem common enough in the US, and Departments of Political Sciences seem to appear in British universities.

  5. beslayed said,

    May 31, 2012 @ 1:18 am

    @Aaron: Of course neither "maths" nor "mathematics" trigger plural verb-agreement….

  6. mollymooly said,

    May 31, 2012 @ 1:18 am

    @Aaron: "maths" is singular. For a singular–plural difference, try "accommodation(s)".

  7. Richard Wein said,

    May 31, 2012 @ 1:55 am

    While I suspect Eric is right to be skeptical about the significance of Kitcher's linguistic point (but I don't know enough French or German to speak with confidence) I feel obliged to quibble with this particular part of his post:

    "…but why would anyone want to elevate such an (entirely hypothetical) asymmetry to the level of facts about the languages themselves?"

    Facts about the way languages are spoken (including facts about the frequency with which words are used) are indeed facts about the languages themselves. Unless Eric thinks that languages have some sort of platonic or essential existence, independent of their usage, (and I doubt he thinks that) what other sorts of facts about a language can there be?

    [ Point taken. My intent was to distinguish between use and structure: it may well be the case (though I strongly suspect that it's not) that French and/or German speakers use the plural more than English speakers do (or whatever the relevant factoid is supposed to be), but it's certainly not the case that there is something inherent to the structures of these languages that dictates or even encourages such differences in usage.

    I do (gasp) subscribe to the idea that use and structure are separable (albeit also non-trivially interdependent in both obvious and non-obvious ways). For example, I take it that it is a (current) fact about the structures of French and German (but not English) that nouns are classified into different genders, and that French has two such genders while German has three; if Kitcher had made some claim about "science" being classified as feminine in French and German, I might have grumbled about any conclusions drawn from this claim but I wouldn't have disputed the claim (or posted about it on Language Log). — EB ]

    Perhaps what Eric really meant to ask is why anyone would want to elevate a fact about a word (the word "science") to a fact about the referent of the word (science itself). As a general response to that question I would point to Wittgenstein's aphorism that "meaning is use". In a discussion of a concept (especially a philosophical discussion) we often need to think carefully about the word's meaning, and its meaning is determined ultimately by its usage.

    [ I don't disagree with this point. If Kitcher's claim is about actual usage and if that claim turns out to be true, then there might be something to discuss here (though again I might grumble about any conclusions drawn from that claim; for example, I might wonder how one would accommodate opposing facts like those that e.g. Adam points out here). — EB ]

  8. Alacritas said,

    May 31, 2012 @ 2:37 am

    Very interesting topic. However I must admit my question is unrelated to the topic at hand, but rather to the syntax of one of the sentences in the article.

    He writes: "*Anyone* interested in this kind of topic (as I am) *are* encouraged to read this essay […]" (asterisks added by me for emphasis)

    My question is why the author has plural verb agreement with 'anyone' here. @Eric, is that how you normally say it? i.e. "Anyone who want to go to the store are invited!" Or was it a typo?

    [ Alas, that was an example of a sentence that was rewritten at the last minute; I think the original subject was "Any Language Log readers". It's been corrected. — EB ]

  9. Ezr said,

    May 31, 2012 @ 2:50 am

    Personally, I find it hard to take Gutting's NYT articles seriously after a couple of disappointing (in the sense of being oversimplistic and almost naive) texts on language issues.
    By the way, it is astonishing how everybody thinks they are entitled to make sweeping claims on linguistic matters, even more technical ones (I can't imagine a non-specialist making equivalent claims on chemistry, for example). Recent posts on this blog have done a great job at pointing out some conspicuous examples of that attitude.

    At any rate, I also happen to disagree with claims of a sharp methodological divide between the "Wissenschaften". I think this reeks of what Chomsky once termed a "pernicious dualism in science" (he was referring to cognitive science, but we mught as well extend the same arguments to "social sciences".)

  10. jayarava said,

    May 31, 2012 @ 3:14 am

    So what is the s doing on the end of mathematics and physics?

    According to the Online Etymology Dictionary the ending -ics "represents a 16c. revival of the classical custom of using the neuter plural of adjectives with -ikos… to mean 'matters relevant to'".

    The original is a Greek adjective: 'mathematike tekhne' (sg). And ca. 14th C we borrowed Latin 'mathematica' (pl) to give us 'mathematic' (sg). Then ca early 17th C we added the s to give us mathematics. (pl form used as sg). The US shortened this to 'math' ca. 1890; which the Brits attest to 'maths' in 1911.

    Meanwhile in the 1930's we Antipodeans invented the phrase 'blinded with/by science'. And what the fellow above has done is 'blind people with linguistic science'.

  11. Ben Hemmens said,

    May 31, 2012 @ 3:20 am

    Gabriele is right. Wissenschaft refers to just about every form of organized learning. Even Rudolf Steiner was able to call a book Die Erziehung des Kindes vom Gesichtspunkte der Geisteswissenschaft; his Geisteswissenschaft is of course his own brand of mysticism, but there's no big semantic strain in this use of the word; Wissenschaft does not really imply any particular kind of epistemological background to the knowledge in question. If you want to say something equivalent to "science", you have to say Naturwissenschaft; and even so, because of the different histories of science over the last 200 years or so in the English- and German-speaking countries, Naturwissenschaft still has a different cultural flavor than "science".

  12. Adam said,

    May 31, 2012 @ 3:46 am

    Hmm, I wonder what deep implications for the philosophy of science arise from the difference between English physics and French physique or German Physik?

  13. Richard Wein said,

    May 31, 2012 @ 3:59 am

    Hi Gabriele,

    Most English readers too will have a hard time understanding why anyone feels a need to explain why "history is also a form of knowledge". Kitcher's concern with that question is presumably prompted by the fact that some people claim science to be the only source (or only reliable source) of knowledge. Since Kitcher probably considers history to be outside science he reads these people as claiming that history is not a source of knowledge, or perhaps thinks that such a conclusion is entailed by their beliefs, whether they actually claim it or not. But he fails to notice that those who say that science is the only source of knowledge typically define "science" in such a broad way that all good empirical thinking (including that of historians) becomes "science". They thereby avoid the conclusion that history is not a source of knowledge.

    Many people (and I would be one of them) would object to that view on the grounds that it misdefines the word "science". I don't think it's helpful to call history a "science" (and still less to call the softer humanities "sciences"). In this regard your point about the German use of 'Wissenschaften' is very interesting. It seems that "Wissenschaft" is more broadly applicable in German than "science" is in English. Applying the word "Wissenschaft" to history is apparently much less controversial than applying the word "science" to it. That would make it difficult to translate this controversy from English into German.

    I don't say this as a criticism of Kitcher's reference to French and German. I think that reference was made in support of a narrower point, namely that there are significant differences between sciences. Regardless of whether the French and German use of the plural supports that point, I consider it to be largely a red herring. The less the "sciences" have in common the less grounds there are for dividing fields of study between "science" and "not science". That is perhaps the point Kitcher is making. But it's also the point made by many of those accused of "scientism"!

    The main problem with Kitcher's article is that he makes no effort to establish the views of the people who are accused of "scientism". Many (probably most) of those people hold views rather different from the ones Kitcher criticises. The trouble is that people only rarely use the word "scientism" to label their own views. It is generally a pejorative word used to label other people's views. The labellers often don't understand the views they are labelling, and in any case the label gets applied to a wide variety of different views. Kitcher seems to assume it's obvious what "scientism" means, but that's far from the case.

  14. Dick Margulis said,

    May 31, 2012 @ 5:25 am

    Was "anyone … are encouraged" a test of the reader's sensitivity to agreement in the wild? Or just a mildly interesting error?

    [ See above. — EB ]

  15. Army1987 said,

    May 31, 2012 @ 6:17 am

    Lots of people in Italian say droga (lit. ‘drug SG’) as a non-count noun for all kinds of illegal psychoactive substances. And my impression is that the kind of people who do that does largely coincide with the kind of people who don't think there's any important qualitative difference between cannabis and heroin.

  16. Vicki said,

    May 31, 2012 @ 6:48 am

    We say "computer science" and "materials science" but "the sciences," and when talking about a lot of science may not use the word at all. We say "biology" or "physics" or "organic chemistry." And don't say "chemistries" even though there's organic and inorganic. "Earth science," grammatically singular, includes geology, meteorology, and sometimes astronomy. Those references to computer science, materials science, earth science, even life science may be skewing Kitcher's dataset.

    Come to think, "history" is usually singular, but we talk about "case histories" (medical) and "oral histories" (part of broader "history" as an enterprise).

    Tangential: I was in a discussion elsenet a couple of days ago about Patricia Cornwell, in which I pointed out that she seemed to be completely ignoring and undermining scientific techniques in her book about Jack the Ripper. Because the lines are fuzzy: history is not a science, but if a historian starts out with "maybe this happened" and then does research in archives which show that it didn't, s/he should say "okay, I was wrong, what did happen?" rather than ignoring the results of your research.

  17. Ellen K. said,

    May 31, 2012 @ 7:05 am

    Isn't it actually, when used as "science" with no article, a non-count noun?

    [ Yes, but (a) it is still grammatically singular and (b) I think Kitcher's argument that this use might encourage English speakers to think of "science" as a singular, undifferentiated mass is unaffected either way. — EB ]

  18. andrewD said,

    May 31, 2012 @ 7:12 am

    @Richard Wein One of the targets for scientism attacks is Jerry Coyne( at "Why Evolution is true"). Jerry has an exchange with Kitcher here:- which language Log readers may like to look at

    [ And if said Language Log readers had checked out the first two links in my post, they would have been led to the same place. — EB ]

  19. Andy Averill said,

    May 31, 2012 @ 7:23 am

    English uses "sciences" all the time, although usually with a modifier. Google yields references to life sciences, physical sciences, meteorological sciences, health sciences, mathematical sciences, etc.

  20. Nelson said,

    May 31, 2012 @ 7:29 am

    @Richard, I spent most of my undergrad years living in a house with a bunch of physicists who very firmly held 'scientist' beliefs of the sort Kitcher is responding to. I've also encountered the same thing repeatedly from undergraduate students of chemistry and math. The postgrads I've associated with since then tend to be much more broad-minded, so maybe you can chalk a lot of 'scientism' up to simple immaturity – but that viewpoint definitely exists in a strong form, and seems to be reasonably common.

  21. Pete said,

    May 31, 2012 @ 7:50 am

    @Army1987 – the English word drugs is syntactically plural, but semantically it can be used as a normal plural ("Heroin and cannabis are two different drugs") or with the meaning of a mass noun ("I'm not doing any drugs tonight").

    The mass-noun usage is still syntactically plural but it can apply equally to a single substance. For example if someone who'd taken only cocaine (i.e. one drug) was asked "Have you taken any drugs?" answered "No" that would be considered a lie.

    The singular form drug is thus only used to mean "type of drug". So I'm not sure there's that much of a difference in the usage of the English and Italian words.

  22. Roger Lustig said,

    May 31, 2012 @ 8:23 am

    Die Wissenschaft hat festgestellt, festgestellt, festgestellt,
    Dass Coca-Cola Fett enthält, Fett enthält, Fett enthält.
    Drum trinken wir auf jeder Reise, jeder Reise, jeder Reise Coca-Cola eimerweise…

    …the German nonsense song that came to mind within a second or so of seeing Kitcher's nonsensical assertion. Here "Die Wissenschaft hat festgestellt, dass…" means precisely, utterly the exact same (heh) as "Science has determined that…"

  23. Dan T. said,

    May 31, 2012 @ 9:05 am

    When somebody is prosecuted for a drug-related offense, American usage calls it a "drug offense", but I think the British say "drugs offense".

    On the other hand, Americans refer to "sports" (even when talking about only one sport; we can say "I like sports; I go to lots of baseball games"), but I think the Brits just say "sport" (as in a newspaper's "sports section" vs. "sport section").

    "Comics" is normally plural when referring to the medium of comic books or strips, but you can also refer to a single strip or issue as a "comic".

  24. Dan T. said,

    May 31, 2012 @ 9:06 am

    Actually, in Brit English, that would be "drugs offence".. gotta get the spelling right.

  25. MBM said,

    May 31, 2012 @ 9:30 am

    Unfortunately, it is not uncommon for otherwise reasonable people to support their otherwise reasonable arguments with factually incorrect linguistic claims.

  26. Ruben Polo-Sherk said,

    May 31, 2012 @ 9:32 am

    In a sense, it doesn't really matter whether the statement that French and German tend to use a plural while English uses the singular is true or not–from the perspective of linguistics, there's a deeper level on which it is wrong, and that is that it's the bad kind of linguistic relativism, the one that gets it backwards and says the way ideas are compartmentalized in a language decides how they are compartmentalized in your head.

    [ When stated in its strong form ("language decides"), I agree. But weaker forms of linguistic relativism ("language influences") aren't so obviously false, though just conjecturing about them in the way that Kitcher does is not any way to go about doing the research necessary to show that influence. — EB ]

    Anyway, the statement is really just a transitional device and not really an argument, and so the deep wrongness is irrelevant to the content of the essay.

    [ Agreed, which is why I noted at the end of my post that it could have just been left out. — EB ]

    I also have difficulty understanding the line about "facts about the languages themselves" that Richard brought up. Does it mean "it's illogical to say that this asymmetry reflects a difference between a 'German way of thinking' and an 'English way of thinking'"?

  27. Mr Fnortner said,

    May 31, 2012 @ 9:57 am

    When Eric says "elevate such an (entirely hypothetical) asymmetry to the level of facts" I believe he means "facts" in the sense of "truths". And I believe he is saying that "such an (entirely hypothetical) asymmetry" is not a truth, not a fact, and that he is questioning why the writer would assert such a thing.

    Furthermore, people make such assertions because they can get away with it with their larger audience, though not with linguists. As a conjecture, I suppose that most people have no knowledge with which to evaluate such an assertion and (would prefer to) simply rely on the statements of an apparent expert.

  28. RP said,

    May 31, 2012 @ 10:08 am

    (You wrote: 'if someone who'd taken only cocaine (i.e. one drug) was asked "Have you taken any drugs?" answered "No" that would be considered a lie.')

    I don't think there is anything remarkable about the word "drugs" in this respect. If I asked "Have you eaten any doughnuts?" and you answered "No" because you had eaten only one, or only one kind, or only one of only one kind, that would be considered a lie too.

    @Dan T.,
    I have heard that, too, but as a Brit I can't be sure whether I would actually say "sport section" or "sports section". Perhaps I never need to refer to it, so the question doesn't arise. But doesn't "sports section" sound the same as "sport section", at least in casual conversation? If so, the difference might turn up mainly in writing.

    In Britain we also use the term "comic" to refer to what you call a "comic book". As far as I know, the term "comic book" is not much used in Britain, except among enthusiasts for the American variety.

  29. QET said,

    May 31, 2012 @ 11:57 am

    Regarding the substance: this is a very old debate. I recommend Gadamer's Truth and Method and Feyerabend's Against Method. Horkheimer & Adorno's Dialectic of Enlightenment is also good. There are countless others.

    Regarding the syntax: this sounds like just the sort of thing that needs a lengthy treatment via mining operations on Big Data as discussed in another post. I'm sure some data would clear this all up.

  30. peterv said,

    May 31, 2012 @ 3:38 pm

    Mathematics, Astronomy, Physics, Chemistry, Geology, Biology.

    Computer Science, Political Science, Social Science.

    I am reminded of that old joke that any academic discipline with the word "Science" in its name is not one.

  31. Svafa said,

    May 31, 2012 @ 4:47 pm

    Off-topic for the LL post, but on-topic for the post in question: I happen to be reading Kuhn's Structure at the moment. Isn't it interesting how these sorts of things coincide, or at least how we perceive such coincidences.

    @Richard: I'm going to agree with Nelson on its prevalence. I had a similar experience as a humanities major back in college, and still run into it a decade later as a computer programmer. There's some tendency to think physics or math or biology is superior to history or language or art, and that knowledge derived from the former is more meaningful or more true than knowledge derived from the latter.

    I think Kitcher does a decent job summarizing the reasons for this in his five observations, especially the fifth: "there is the commonplace perception that the humanities and social sciences have been dominated, for long periods of their histories, by spectacularly false theories, grand doctrines that enjoy enormous popularity until fashion changes, as their glaring shortcomings are disclosed."

  32. Janes'_kid said,

    May 31, 2012 @ 5:47 pm

    A fine magazine The Sciences was published from 1961 to 2001.

  33. Ruben Polo-Sherk said,

    June 2, 2012 @ 9:33 am

    EB, I don't dispute at all that language can influence thought; but since his claim could only possibly be relevant if it's taken to be the strong form, that's how I expressed my point.

    [ How is the weak form of the claim not relevant here? Kitcher could very well be suggesting that the (over-)use of singular 'science' in English non-negligibly influences the minds of English speakers to think of science as a singular thing, whereas the (over-)use of plural 'sciences/Wissenschaften' in French/German non-negligibly influences the minds of speakers of those languages to think of science as a plurality. Indeed, I can't imagine that he is suggesting anything stronger than that. — EB ]

  34. Ruben Polo-Sherk said,

    June 3, 2012 @ 9:58 am

    Kitcher made the statement in discussing that natural science is able to be precise by looking at a simplified view of reality; one kind of this simplification is that that divides the different fields of science into "the sciences".

    His point would have to be that people (in particular, English speakers) are unaware of this kind of simplification because, due to the word "science" being overly prominent, they don't realize how each field looks at the world from a narrow perspective. And that therefore, they are not aware of the limitations of science–i.e. a reduced capacity to generalize–and so they only see the good.

    Put into the perspective of linguistic relativism, English speakers have a collective cultural habit of not being able to see the narrowness of each field because more salient in their minds is "science", as an all-encompassing field.

    But the only way this can be caused by which form ("science" or "the sciences") is used is if inherent in the concept "science" is the aspect that there is no division between the sciences. So, the only way Kitcher's argument works is if you assume that in English, you have this concept, "science", which doesn't differentiate, and this other one, "the sciences", which does. In that scenario, you're absolutely right: it's the weak form, and I'd have no objection to it on those grounds.

    But, for me, the term "science" doesn't unify in the sense that it would have to for this assumption to be valid. Although that conceptualization is compatible with a singular form much more so than it is with a plural, it's not intrinsic. So, underlying that assumption is the idea that the fact that it's singular forces the concept that it represents to be seen as indivisible.

    Certainly, "the sciences" explicitly differentiates the different fields in a way that "science" cannot. But I don't see in the semantics of "science" a unification that makes the boundaries between fields invisible. For me, it's an abstract generalization of all the things that are science, and so innately the crucial distinctions are present. Perhaps for other people, the term "science" does represent a monolith; there, the weak form perspective would be valid. But I, and probably many other people, have a different conceptualization.

  35. Michael A. Lowry said,

    June 5, 2012 @ 5:52 am

    The German word for science can singular or plural, just as in English. Moreover I doubt there is any correlation between whether a language treats science as a singular or plural and whether social sciences and the humanities are considered serious sciences.

    If you translate the parts of the German word for science directly back to English cognates, you get something like ‘knowship.’ Science, the word tells German speakers, is the process of knowing. If you do not have knowledge, you are ignorant; who wants to ignorant?

    I would therefore be more inclined to believe that because the word for ‘to know‘ is embedded in the word for ‘science,’ the German word has positive connotations that are absent in English.

    Also, as others have pointed out, the words ‘mathematics’ and ‘physics’ were once considered the plural forms of singular words.

  36. John Hadfield said,

    June 9, 2012 @ 8:55 am

    The word "science" is taked from the Latin word "scientia", which means knowledge, derived from the Latin word "scio" = I know. When the word science was first introduced into English (and French, and other Latin languages) the people using that word all had some knowledge of Latin, so they knew what it meant.

    I can't see much difference between that and the use of the German word "Wissenschaft".

  37. Thor Leifsen said,

    June 28, 2012 @ 8:22 pm

    Knowledge can refer to the theoretical or practical. Would not science refer to the theoretical or practical also? Germanic languages divides the theoretical and practical in to,

    Vitenskap – Norwegian
    Wissenschaft – German

    Kunnskap – Norwegian
    Kenntnisse – German

    That the word science/scientia means knowledge is probably not regarded as scientific knowledge. Some cultures think that science should relate to all knowledge as one truth.

RSS feed for comments on this post