Various types of whatever(s)

« previous post | next post »

Joan Maling writes:

The various co-authors on a neurolinguistics paper (I am one)  have different judgments about the following:

a.    two principal kinds of hypothesis  
b.    two principal kinds of hypotheses  

The two British co-authors prefer singular hypothesis; two Americans prefer plural hypotheses.  Curious. Has anyone looked at this variation, either as an idiolectal or a dialectal difference?

I don't know of any relevant research — but it's a big world, so there may well be something Out There.

Joan and her co-authors did a quick web search for the alternatives, and concluded that both of the options are possible.

I decided to extend their experiment a little bit,

(1) using Google Scholar;
(2) eliminating the modifier principal;
(3) generalizing the quantifier two to two|three|four|several|many;
(4) generalizing kinds to types|kinds|sorts; and
(4) generalizing  hypothesis to hypothesis|answer|idea.

This focuses on the register of scientific and technical writing, and produces numbers for each sub-search that are (usually) large enough to be meaningful, but small enough to check whether Google's cavalier "About N results" estimate is close to being valid.

The results are consistent, to an extent that I found surprising. It's true that both options are Out There, but overall, authors in this type of writing use the plural form of the of-phrase noun about three times in four. Thus Google Scholar gave me 369 hits for "two types of answer", and 1030 hits for "two types of answers", where 1030/(1030+369) = 0.7362402, i.e. about 74% for the plural option. And so on:

hypothesis/es answer/s idea/s MEAN
two types of 439 ~ 921 (67%) 369 ~ 1030 (74%) 42 ~ 316 (88%) 72.7%
two kinds of 199 ~ 423 (68%) 300 ~ 763 (72%) 59 ~ 462 (91%) 74.7%
two sorts of 17 ~ 29 (63%) 71 ~ 248 (78%) 13 ~ 161 (93%) 81.3%
MEAN 67.7% 73.4% 89.2% 74.3%


hypothesis/es answer/s idea/s MEAN
three types of 147 ~ 342 (70%) 158 ~ 550 (78%) 25 ~ 140 (85%) 76.3%
three kinds of 35 ~ 137 (78%) 88 ~ 271 (75%) 17 ~ 119 (88%) 79.0%
three sorts of 3 ~ 15 (83%) 13 ~ 63 (83%) 1 ~ 41 (98%) 87.5%
MEAN 72.8% 77.3% 87.5% 77.5%


hypothesis|answer|idea hypotheses|answers|ideas percent plural
four types|kinds|sorts 150 464 75.6%
several types|kinds|sorts 103 349 77.2%
many types|kinds|sorts 71 354 83.3%

Some of the singular examples are disguised plurals, as in "the two types of answer sheets" — that presumably means that these numbers slightly underestimate the plural preference.

There's a suggestion of further structure, with an ordering of plural preference

hypothesis << answer << idea


type << kind << sort

I didn't try to check for geographical or temporal (co-)variation — perhaps some enterprising readers will want to take this little experiment further.

This is the kind of thing where I can imagine some self-appointed usage authority deciding that one of the two alternatives is illogical and therefore unacceptable. And there are people who believe that allowing two valid answers to any grammatical question is like playing tennis without a net. But as far as I know, this opportunity to create a Rule is still unclaimed.

Update — following Matt_M's suggestion, some Google ngrams evidence for an emergent British-vs.-American difference. First the British corpus:

And then the American one:


  1. Robot Therapist said,

    August 27, 2015 @ 6:09 pm

    Hmm, that's interesting.

    My reaction (I am British) was "hypothesis", which agrees with your introductory remarks.

    However, I believe I would say "two types of oranges" and "two types of cars", so I wonder if it is something about "hypothesis". Perhaps that it's an abstract noun, or that its plural is an unusual word. Or perhaps that in talking about "two types of hypothesis" we are clearly indulging in some highly academic thinking, and therefore more formal usage is called for?

    [(myl) An idea that occurred to me: perhaps the singular "two types of X" is implicitly treating X as a sort of mass noun. But then I observed that when the type/kind/sort word is singular, and a single X is under consideration, then X must be singular as well:

    That's the type of animal it is.
    *That's the type of animals it is.

    But even with singular "type of X", you can get plural "type of Xs" when more than one are in the picture:

    Given the type of animal they are, …
    Given the type of animals they are, …

    So it's apparently not just a covert mass/count distinction.

  2. Joseph Pentheroudakis said,

    August 27, 2015 @ 6:28 pm

    Some inconclusive discussion by the masses:

  3. Ethan said,

    August 27, 2015 @ 6:44 pm

    @Robot: I think your oranges/cars examples bring out another dimension. I think I would say "They are selling two types of oranges at the store" but "Valencia and Fukuhara are my favorite two types of orange". In the former case there are clearly multiple oranges involved. In the latter case I may have in mind a selection that results in my having a single orange. The same issue may come into play with hypothesis/hypotheses, depending on the context. In the original example "principal kinds of …" signals that a lot of something are involved, so I think I would use the plural "hypotheses". On the other hand I don't think I would ever say "A and B are different kinds of hypotheses".

  4. mike said,

    August 27, 2015 @ 7:27 pm

    Classic programming/nerd humor: "There are only 10 types of people in the world: Those who understand binary and those who don't."

    So, another vote for plural.

    PS Get the t-shirt:

  5. Mara K said,

    August 27, 2015 @ 7:30 pm

    Robot Therapist: the sense I get from "hypothesis" is that when you're saying or reading it quickly, it looks like it's already plural. What are your judgments on answer/answers or idea/ideas? I'm from the US and choose plural for all of them, but have to slow down to think about hypothesis/hypotheses.

  6. David L said,

    August 27, 2015 @ 7:32 pm

    As a Brit turned American, I have long considered this an example of a transatlantic difference with no obvious explanation. British English prefers "types of thing" and "sorts of thing," American English prefers "types of things" and "sorts of things." I don't think the nature of the noun makes any difference.

    I have no evidence for this, just general observation over the years.

  7. James said,

    August 27, 2015 @ 7:34 pm

    I'm American, but I prefer the singular. But the plural version doesn't sound wrong to me; I just wouldn't use it myself.

  8. Mara K said,

    August 27, 2015 @ 7:34 pm

    David, out of curiosity: how often do you catch Americans using "sorts of things" as opposed to "types" or "kinds of things"?

  9. Rubrick said,

    August 27, 2015 @ 7:58 pm

    Blunted by mike's comment, but:

    There are two kinds of people in this world: those who would say "There are two kinds of people in this world: those who would say "There are two kinds of people in this world….

    Sorry, "two kinds of person" was supposed to be in there somewhere but something went horribly wrong….

  10. Brett said,

    August 27, 2015 @ 8:03 pm

    With the singular "hypothesis," it immediately sounded to me like a Britishism.

  11. Rubrick said,

    August 27, 2015 @ 8:07 pm

    Okay, the meta-est thing EVAR just happened when I tried to comment on this post.

    First I was going to post a simple joke (blunted by mike's) about two kinds of people, those who said "two kinds of people" and those who said "two kinds of person".

    Then I realized it was much more interesting to refer to two kinds of people, those who would say that joke one way and those who would say it the other, which made the whole thing recursive, with a "Those who would say: "Those who would say:"… sort of structure (I'm not quoting that version directly for reasons that will become clear.)

    After bailing from the recursion with an ellipsis, I commented that something had gone horribly wrong.

    Then I clicked "Submit Comment", and was told by WordPress that it was a duplicate comment which I had already submitted (I had not)! Apparently the repeated lines had somehow tricked some badly-written WordPress algorithm into thinking the comment, once posted, would be a duplicate (and so it refused to post it).

    Which means something *did* go horribly wrong.

  12. John Roth said,

    August 27, 2015 @ 8:57 pm

    Interesting question. I find "two types of car" correct, but "two crates of oranges" is also correct; "two crates of orange" simply sounds wrong.

    I think the difference is that a crate contains multiple oranges, so it would be "one crate of oranges," and the number of crates doesn't matter, while words like "types" and "kinds" requires a singular, at least for me. It's distinguishing parts of an umbrella term, if that makes any sense.

  13. Yuval said,

    August 27, 2015 @ 9:14 pm

    The UK/US discrepancy probably shares a root cause with the different treatment of collective nouns like "government" and "committee", no?

  14. Matt said,

    August 27, 2015 @ 10:32 pm

    To me there is a subtle semantic distinction. The singular "hypothesis" evokes the sense of a prototype/abstraction, whereas the plural "hypotheses" seems to refer to a set of specific entities (or at least it introduces such a set into the discourse in a way that the singular form does not).

    Perhaps this is a definiteness difference, or perhaps there is indeed a distinct sense of the word "hypothesis" that refers to the (Platonic Form) ideal of the entity, which is necessarily singular. This may be a general pattern across most nouns, e.g. with "orange" potentially referring to the idealized concept of what it is to be an orange.

    When I say that there are "two types of orange," I'm definitely not referring to any specific collection of oranges at hand. Rather, I'm referring to the class of all oranges. In contrast, the phrase "two types of oranges" could refer to a specific collection of oranges, e.g. sitting before us in a crate.

  15. michael ramscar said,

    August 27, 2015 @ 11:40 pm

    meanwhile, in other semantic/usage splitting news, 'neurolinguistics' slips by unnoticed to win by a head.

  16. Martin Ball said,

    August 28, 2015 @ 12:17 am

    @Brett: With the plural "hypotheses," it immediately sounded to me like an Americanism.

  17. Doreen said,

    August 28, 2015 @ 12:39 am

    As an American turned Brit, I have long had the same impression as David L of this as a UK/US difference but haven't done any systematic study of it.

  18. Keith said,

    August 28, 2015 @ 2:57 am

    I immediately thought "two principal kinds of hypothesis" would be my natural choice.

    To test, I also re-wrote the phrase, but as "two breeds of dog" and "two varieties of apple", to reduce the number of words and to use much more common nouns.

  19. Simon said,

    August 28, 2015 @ 3:05 am

    I'd go for hypothesis in the sentence as presented. However hypotheses would be the one to go for if the sentence were changed round to something like: Hypotheses can be grouped into two types…

    I'm british, but have a eighties comprehensive school education, so you may be best to default to the opposite of my opinion.

  20. mollymooly said,

    August 28, 2015 @ 3:06 am

    Somewhat related is "these kind of…", which is deprecated but Out There.

  21. Bob Ladd said,

    August 28, 2015 @ 3:15 am

    Weird. As another American turned Brit, I think I'm reasonably sensitive to [this kind of thing]/[these kinds of things]/*[this kind of things]/??[these kinds of thing], but in 30+ years on this side of the Atlantic I had never noticed that there might be a Brit/Am difference here.

  22. Matt_M said,

    August 28, 2015 @ 4:01 am

    Google ngrams confirms that there's a big difference between American and British usage with regard to pluralisation of the noun after "types of":

    American English
    British English

  23. Michael said,

    August 28, 2015 @ 5:30 am

    And how is this related (or is it) to the ambiguity in "their life" vs. "their lives"? "their wife" or "their wives"?

  24. Eric P Smith said,

    August 28, 2015 @ 5:33 am

    @Bob Ladd: To my ears, "These kind of…" and "this kind of things" are definitely ungrammatical; "these kinds of things" and "these kinds of thing" are all right, but only if we are indeed concerned with two or more kinds; while "this kind of thing" is fine. But if I am talking about things and not kinds, which is generally the case, I prefer "things of this kind".

  25. Andris said,

    August 28, 2015 @ 5:52 am

    Perhaps I've missed something in the above, but isn't it a simple question of countable/uncountable nouns?

    two principal kinds of hypotheses
    two principal kinds of technology
    two principal kinds of beer
    two principal kinds of beers? (certainly less common in BrEng: perhaps more specific)

    Hypothesis is grammatically singular…. (with hints of uncountable abstraction)
    and as Matt says the singular suggests a sense of a abstraction, whereas the plural seems to refer to a set of specific entities

  26. David L said,

    August 28, 2015 @ 8:20 am

    @Mara K: I can't give you any quantitative analysis, alas! There may indeed be a difference between 'types' and 'sorts.' I also have the (entirely qualitative) impression that Americans are less likely than Brits to say "sorts of things" rather than "types of things."

    As I say, this is just my general feeling based on many years of casual observation.

  27. Y said,

    August 28, 2015 @ 12:22 pm

    FWIW, "there are two kinds of people" seems much more acceptable than "there are two kinds of person" or even "there are two kinds of persons".

  28. Martin Keegan said,

    August 29, 2015 @ 6:13 am

    There's someone from Düsseldorf doing some sort of survey on Reddit on this question right now (possibly because they saw your post); this was about the word "type" rather than "kind" but it's analogous.

RSS feed for comments on this post