A Wildly Popular Verb

Back when I was interviewing for jobs (so this would have been during the last few days of the year 1989 in Atlanta), my writing sample was “Contextualism and Knowledge Attributions” (which later appeared in PPR), and a quite famous* philosopher insisted to me during an interview that it was in fact extremely uncommon for people to say that they do or do not “know” things, and that I only thought otherwise because I spent all my time talking with other philosophers. I remember him exclaiming, “Why, I go whole weeks without using that word!” I tried to convince him otherwise (well, not about his own usage, which I couldn’t really speak to, but about his general point), citing all the various kinds of situations in which it is utterly natural to use “know(s)”. He wasn’t budging: “No, I almost never hear that.” I in fact spent a whole lot of time with non-philosophers, and I thought this guy was, well, full of shit, as my non-philosophical friends liked to say (almost as much as they liked to say “know(s)”!)

Has anybody else heard such claims — about “know(s)” being seldom used in non-philosophical talks — from philosophers? Do any CD readers think this themselves? Does anyone have any insight into what could be behind such claims? I have heard it from time to time since then.

A few days ago, another quite famous philosopher told me the same thing again. This time, I thought to check it out on Google. (Of course, I couldn’t do that in 1989, but could have done it for a few years now.) Turns out, at least according to Google, it is in fact wildly popular — one of the most commonly used verbs in the English language, at least so far as I can tell. I’m not an expert at Googling, but I did a search on “know OR knows OR knew”, and compared it with some other results. Here’s what I get right now:

know OR knows OR knew: 2,060,000,000

give OR gives OR gave: 1,680,000,000

play OR plays OR played: 1,410,000,000

remember OR remembers OR remembered: 688,000,000

The trouncing of “play” is especially impressive, because “play” and “plays” are also a singular and plural noun — and a quite common noun on the internet, since sports are all over the web (double play, what a great play, etc.).

Well, you get the idea, and can now check for yourselves. In fact, you may be much better at this than I am. I’ve heard that somewhere one can find a list of the most commonly used English words — and that “know(s)” ranks pretty high. But I haven’t yet seen such a list.

Of course, not all of these results are uses in which it’s said someone knows that something is the case (propositional knowledge), but, when you dip into the results a bit, it seems that a whole lot of them are.

So, if you are ever worried that “know(s)” is mostly just philosopher-speak, it looks to me as if that’s wrong.

*by “famous,” I just mean “philosophically famous.” As I explain to my students when I catch myself calling a philosopher or a philosophical position or argument “famous,” I suspect that the most famous American philosopher, whoever that might be, is much less famous in America than is the least famous of the Spice Girls (Sporty, perhaps?), though the Spice Girls were British, broke up quite a few years ago, and never did anything worth mentioning.


Comments

A Wildly Popular Verb — 40 Comments

  1. I’ve heard philosophers claim that ordinary usages of ‘knows’ are radically different from what most philosophers assume they are – but I haven’t heard philosophers say that it’s usage is uncommon.

    A couple of asides on very popular uses of ‘knows’ in pop culture.

    First: Hasn’t this series of public service announcements been around forever – http://www.themoreyouknow.com/

    Second: At the end of every G.I. Joe cartoon. The format was always the same. (a) kids were doing something bad, (b) GI Joe character would come along and educate them, (c) this slogan was stated “Now you know, and knowing is half the battle”

    Not sure if that shows much.

    BTW-I’m typing this while watching a morning news show and the anchor just said “We still don’t know…”

  2. I hear ‘I don’t know’ used by non-philosophers all the time. Your interviewer doesn’t _know_ what they’re talking about.

    A note on Google. Results reflect writing on web sites, not spontaneous speech. Ask a linguist–they’ll have a better read than Google. Maybe someone like David Crystal would _know_.

    ST

  3. My guess is, he was referring to a technical interpretation of “know”, not the usage of the actual word. (However I don’t know what his interpretaion of “know” would have been.)

  4. No, this was some kind of objection to doing ordinary language philosophy on “know(s)” on the grounds that there just isn’t much ordinary usage to account for, since it’s such a very rarely used word in non-philosophical settings. (My paper certainly doesn’t give the impression that I was seeking to discuss some technical use. It opened with some cases that I thought displayed an important feature of the ordinary use of “know(s).” He thought that was crazy because, outside of philosophy, “nobody really talks like that.”) When he said he went whole weeks without using the word, he didn’t seem to be referring only to some special, technical use.

  5. I agree that “knowledge” talk is utterly natural.

    FWIW, in my experience young children speak of what they and others “know” and “don’t know” before they speak of “believing,” “thinking,” “seeming,” etc.

    But I do happen to disagree with Keith when he says, “the most famous American philosopher, whoever that might be, is much less famous in America than is the least famous of the Spice Girls.” I bet Thomas Jefferson is the most famous American philosopher, and he’s more famous here than Sporty et al!

  6. The idea that a page count from searching ‘know OR knows OR knew’ will give you an estimate of the frequency of the use of the verb ‘know’ is spurious. The first search term, ‘know’, would also yield pages including ‘knowledge’, the noun form.

    This is not to dispute the conclusion; just the method.

  7. Those who know (!) more about Googling can correct me if I’m the one who’s wrong here, but I *think* you’re wrong about how these searches work, Gregory: A search for ‘know’ doesn’t count instances of ‘knowledge’. This seems confirmed by this little experiment: I just searched ‘know’ by itself, and got 1,680,000,000 as the result; I then searched ‘know OR knowledge’ and got 2,070,000,000 results; a search for ‘knowledge’ by itself got 432,000,000 — pretty much what one would expect if a search for ‘know’ does NOT count pages just because they include the word ‘knowledge’, when we keep in mind that some pages will include both ‘knowledge’ and ‘know’. Then I re-searched ‘know OR knows OR knew’ and this time got 2,070,000,000, and then added ‘knowledge’, searching for ‘know OR knows OR knew OR knowledge’, and got about a ‘knowledge’s worth of extra hits, with the result of 2,430,000,000. Why would I get those extra hits if the first search was already counting instances of ‘knowledge’?

    If this is “spurious,” it certainly doesn’t seem to be so for the reason you give. The worry in comment #3 is a genuine one. But this little adventure in Google-land still seems to be providing pretty good evidence for the conclusion that “know(s)” is an exceptionally popular verb.

  8. John is right that children use “knows” earlier and more heavily than “thinks”. It’s the main verb in about 70% of mental state ascriptions, more than twice the incidence of “thinks” (and “believes” is really rare among kids–less than 4%).

    That’s from Bartsch and Wellman’s longitudinal studies of kids up to age 6, working from coded transcripts of 200,000 spontaneous child utterances. The coding is important: they went through and cut out all the formulaic uses which they took not really to reflect mental state ascription (like “You know what?” as an attention-getter). They also cut repetitions. Those uses have some chance of skewing your data — Bartsch and Wellman thought that only 40% of children’s utterances of terms in the think/know/believe group involved genuine reference to thinking and knowing. Even so, ascriptions of knowledge are really common, occurring in about 3.5% of all the utterances those kids were making. I don’t know offhand of any good adult data on just that point, but I’d be curious to hear of it if anyone knows.

    If you wanted to know the oral vs. written data on use of verbs of thinking and knowing you could check the LDC (Linguistic Data Consortium) speech databases, but there is the problem of coding for both oral and written production.

    Incidentally the dominance of “knows” over “thinks” looks to be cross-culturally robust. Twila Tardif says it is also more than double for Chinese-speaking kids.

    So Keith, yes, you are right, and not just about English, and some famous philosopher out there is wrong.

  9. Does anyone know whether ‘knows’ in chinese is easier for chinese speaking kids to pronounce than ‘thinks’ or ‘believes’ in chinese (in the same way that ‘knows’ is easier to to pronounce for english speaking kids than ‘thinks’ or ‘believes’)?

  10. Here’s an interesting debate, given the famous philosopher’s assertion that ‘know’ does not occur in ordinary language: If I told you that on a casual LexisNexis search, the President of the Gosh Darn United States has used the phrases “I know” and “we know” so many times, during his presidency, that one must edit one’s search quite a bit to get the count under 3,000 results, is that proof that Bush is EXTRA-ordinary?, or that the most astoundingly ordinary mind has resorted to saying this phrase for many years? Examples range from the very recent, “I know many Americans have reservations,” re the bailouts, to his first Gridiron club dinner in 2001, at which he said, “I know the human being and fish can coexist peacefully.”

  11. These aren’t the numbers that I have, but set that aside. The searches in general do not respect parts of speech. You need to count terms in a representative corpus, like the Oxford corpus, which does support your claim. Web pages are not necessarily representative of English usage, so I would question your sample. The Google search algorithm does not simply count occurrences of terms, so I would question your sampling method. Finally, the searches return an estimated number of relevant pages, not estimated number of occurrences of the terms, and those estimates are rounded to the nearest million pages!

    So, I’d say the reasoning is a total disaster. But we’re agreed on the conclusion!

  12. I had the same worries as Greg. But, here’s a nice corpus to try:

    http://www.americancorpus.org/

    My searches in this corpus revealed the following frequencies:

    “know”: 599613
    “play”: 105425
    “give”: 141725

    “know*”: 806901
    “play*”: 329186
    “give*”: 273257

    These more salient frequencies clearly support Keith’s thesis (which seemed true to me anyhow).

  13. Great thanks to Jennifer and Brandon!

    I do still think the Google results I cited provided reasonably good grounds for thinking ‘know(s)’ is a very commonly used word. (I would be worried if Google searches for ‘know’ counted web pages as hits because they contained the noun ‘knowledge’, but that really seems not to be the case.) I’ve seen similar results cited as reasonably good reasons for similar conclusions in a very wide variety of settings. The idea seems to be that you likely won’t go too far wrong with such a method, despite how extremely quick and easy it is to apply (so long as you’re on-line). That what look to be better methods lead to quite similar conclusions in this case seem to support that basic idea. I doubt this was just luck!

    In any case, it does seem to be turning out that ‘know(s)’ is extremely commonly used. The suggestion that it is very rarely used always seemed crazy to me. But I don’t think I ever supposed it was as commonly used as it seems to be turning out to be.

  14. Keith,

    The Google search algorithm is designed to deliver relevant pages given a particular query, and it uses the parts of speech of your query as only one of very many factors. So, you cannot depend on it to return pages that respect the parts of speech of your query. There would be no business model for Google, or any other company doing search, if the algorithm reliably worked as you have conjectured. (You can intervene some by placing quotation marks around search terms, but I would not bet that even this would force results to respect parts of speech for very common terms).

    The numeric estimates associated with these queries are so large as to be (nearly) meaningless. Google is in effect telling you that it estimates that there are lots and lots and lots of pages relevant to one of your queries, versus merely lots and lots of pages relevant to another. The significant digits associated with the page estimates under discussion are floating between 1 and 10 million.

    So, we are getting estimates of the magnitude of web pages that *might* be relevant to these different queries. Common sense would give you better evidence for the thesis than these search results, with due respect to the philosopher who was yanking your cord twenty years ago.

  15. So I was just getting lucky? I must be damn lucky, since the results keep lining up pretty well with the various rankings of common verbs that I am now finding (and that I’m mostly presuming are based on better methods than Googling). By my own common sense (which I guess isn’t that great on this matter, even though it was better than my interviewer’s), I’d have guessed ‘know(s)’ isn’t nearly as common as the Google searches indicate and as seems now to be the case given better ways of checking.

  16. I’m not sure who’s right about google searches, but you can check out the frequency of ‘knows’ in Shakespeare plays at
    http://www.mta75.org/curriculum/english/Shakes/

    For example, in Hamlet:
    know’st, 2; Know’t, 1; Know, 74; Knowest, 1; Knowing, 2; Knowledge, 1; Known, 8; Knows, 3. In terms of frequency it beats out: Sir, 74; Them, 74; May, 71; Tis, 71; Go, 70; Us, 69; King, 67; Love, 66; Did, 65; Very, 64; Speak, 63; Which, 63; Hath, 62; Then, 62; Why, 62; Must, 61; Thee, 59; Give, 58; Should, 58. That seems pretty significant.

  17. I’m sure one could use a screwdriver and his own head to open a can of beans, too. Now, I’d argue that this is not the best way to eat your beans. I’d even be so bold as to suggest that a screw driver is a lousy can opener, that it is for screwing things down, not for perforating metal. And, amazed that someone would proudly defend this method, I might forget to point out that there is a health risk to using one’s head as a hammer. But I’d have to concede that luck would play little role in his getting cans open.

  18. proudly defend this method

    You misread me. If you have a good can opener handy, then use that, by all means! (And if you really do have a good can opener handy there, please “spill the beans” and let us know what the real story is on this matter. That would seem more productive than just standing there watching the guy struggling with the screwdriver and harping about how bad a method that is.) I’m saying (see esp. #16) that if you have sturdy screwdriver handy, that seems to work pretty well (using it with your hands to get a better analogy), and that that’ll typically work better than trying to open the can by bashing it with a rock, which I’m thinking would be roughly analogous to just using one’s sense of whether and how frequently used a word is used (at least in my own case). Like I said, I’ve seen that screwdriver being used, apparently to decently good effect, by those who know their beans. I remember, for instance, Kai von Fintel, citing a quick Google search on an entry to his blog to support a report that a certain construction or whatever “is quite common” or “very common,” or something like that in a “A quick Google search reveals that this is quite common…” kind of way. I guess he could have worried that it might not well represent conversational speech, about rounding to the nearest million or whatever, or could have instead gone to the drawer to get a really good can opener, or just used his own sense of how common the item was, but he seemed to think that in a blog entry, for his current purposes, this scewdriver would do fine. I’m pretty sure I’ve seen Google being used in a similar way and in a similar spirit on the blog Language Log. And other similar places. (It could be, though, I admit, that those who I’ve seen Googling to reasonably good effect are better at using this screwdriver than I am, with a better idea of which cans to try it on, how exactly to hold it, etc.)
    A bit later: Here is a discussion I’ve just found of some of the pros and cons of such uses of Google, other search engines, and also some “can openers” (fancy corpora). It compares results for several items (including some individual words, “full” and “empty”) across several of these tools. Seeing the variations in the results might start to give one a sense of how much caution is called for, etc. Those interested in using Google for such tasks can find in the archives of Language Log LOTS of blog posts (many of them listed here) concerning various details about how Google works — esp. details relevant to effectively using it as a tool for getting quick estimates of how common words or phrases are.

  19. For those who, like me, are interested in knowing that, I have found the chart at the top of this page, according to which “know” is the third most common verb for controlling a that-clause in conversation, behind “think” and “say”. For me (though this just shows the way in which my sense of this was skewed), the surprise is how close “know” comes to “say” here — and also the extent to which it clobbers “see”, “show”, “find”, “believe”, “feel”, and “suggest”. Interestingly, some other verbs beat it out for controlling that-clauses in academic prose, but here the way the chart is presented (its use of *s instead of just reporting the underlying numbers) seems to greatly limit its usefulness.

  20. I have no quarrel with using search engines for comparing the estimated frequency of written constructions; many non-native speakers of English use Google to check their intuitions about prepositions and other questions about phrases in this matter.

    But this is not at all what you were proposing to do. You need at least two terms, enclosed in quotes, for this to work out, and I wouldn’t place much stock in searches of common verbal phrases, which would yield very high page counts.

    Just as a screwdriver is not a can opener, Google search is not a word-counter-in-spoken-English machine.

  21. OK. I’ll just note that many of the complaints you raised would apply to estimates of the frequency of multiple words inside of quotation marks as well as to those of single words, and that the Language Log folks certainly don’t seem to recoil from using Google on single words for getting an estimate of their frequency.

  22. I was going to add to what’s in #26 that my impression is that the best reason for not using Google on single words, and especially very common single words, while being willing to tolerate its limitations on other frequency estimates, doesn’t have to do with the method being more unreliable for single word estimates, but rather with the existence of a better alternative: There seem to be estimates of word frequencies out there that are based on more careful methods and that are very easily accessible. The wordcount site Brandon links to directly above looks very promising.

  23. No, my complaints do not apply to estimating the frequency of (most) phrases. They apply to the method behind your original argument and the conclusion you wished to derive. The two applications are chalk and cheese.

  24. I think I know what’s going on here …

    In the context of Keith’s original post, he could permissibly assert that the Google results were good evidence for the conclusion that ‘knows’ & company are very commonly used. He could assert this because he “knew” it.

    Then Gregory chimed in with his worry that the Google results are not “necessarily representative of English usage.” He backed up this general worry with specific, nontrivial ways in which the results might go wrong. This skeptical suggestion threatened to enlarge the scope of epistemically relevant worlds, thereby threatening Keith’s “knowledge,” and thereby threatening to prevent him from any longer permissibly asserting that the Google results are good evidence.

    Keith then responded by providing additional evidence, so that he could continue to permissibly assert that the Google results are good evidence, even in this more demanding skeptical context.

    Sound plausible? Anyone?

  25. Sounds quite plausible John.

    This is probably one of the more heated debates I have seen on this blog (it appears a bit testy)…Anyhow, I must admit, that in spite of what the language folks say, it seems a bit odd to suggest that you can make a legitimate inference about the use of ‘knows’ from typing it into a computer search engine. I wonder how it would be received if you tried to publish about this issue and “googling” was all that you cited for your empirical method?

    Nevertheless, that is not to say that googling isn’t a good place to start. But, it seems pretty clear that if you want to make a justified claim about the *whole population* of English language users, you are going to have to do better than googling for your conclusion.

  26. 29: I was thinking primarily about the complaints that “Web pages are not necessarily representative of English usage” and the complaint that Google doesn’t count occurrences of the word/term, but rather the pages on which it occurs.

    31: Citing Google hits in an academic paper is quite different from in a blog post! I’ve very often seen Google hits cited (to good effect, it’s often seemed to me) in the latter context. Some blog posts — and this very post was an example of this — for instance, are of the “Hey, here’s something – Anybody got something better on this to go here?” variety. And, ignoring unhelpful comments (which is probably what I should have done all along), this post worked very well for the purpose it was intended, thanks to some very helpful comments. (I don’t know how many hits Certain Doubts gets; I doubt it’s all that popular. But I dont’ know of any blog that draws better, more helpful comments. That’s on display here.) That’s just what a blog post is for (or at least one of things it can well be for), I’d say. I wouldn’t cite Google hits in a paper I’m sending off to journal (either for word frequency or multi-word constructions). And now I wouldn’t use it again for word frequency — at least for a fairly common word — even in a blog post, not because it’s generally too unreliable for that purpose (even one’s own sense of how common a term is, after all, seems at home in various blog posts), but because I now know of better sources on that matter that are just as easily accessible, thanks to this very discussion. (Well, there could be exceptions: like a word that one has reasons to suspect has only very recently become very popular. There, I can easily see writing a blog post citing a Google count to support the point that the word seems to be becoming quite popular. Some other possible exceptions I won’t bore you with.)

    *************************

    On more substantive matters: Does anybody have any ideas about the contrast between what Jennifer reports in the first paragraph of #9, where “know” beats “think” among children in “mental state ascriptions” and the chart I link to in #24, where “think” seems to more than double-up on “know” in “controlling that-clauses”? Is this likely due primarily to a difference between children and adults, some difference in the exact question being addressed, or some difference in methodology? (Unfortunately, I haven’t been able to find much explanation in the book I link to (CORPUS LINGUISTICS) about how exactly the results in the chart were arrived at. Maybe that’s in a part of the book not available on Google books, or that I just haven’t found yet? But maybe someone knows.)

  27. I suppose the discrepancy between the list of most common words linked to in comment #1, in which ‘know’ is listed as the 59th most commonly used word, and wordcount, which is linked to in comment #27, and at which ‘know’ is listed as the 83rd most common word, is largely due to the first source ranking “lemmas” (in which ‘know’ is ranked as part of a larger team that includes ‘knows’ and ‘knew’), while in the wordcount ranking, ‘know’ and ‘knows’ and ‘knew’ each competes on its own (and come in at # 83, 1222, and 336, respectively; some other ‘know’-related words on wordcount: ‘known’ is 341, ‘knowledge’ is 675, ‘knowledgeable’ is 14497).

    I’m assuming that when wordcount ranks, for instance, ‘play’, which it says is the 433rd most common word, it is including verb and noun uses together in that ranking, and similarly for ‘plays’, which clocks in at #2782? (Which in turn makes me wonder whether the British National Corpus, on which the rankings are based, is a bit light on sports sources and sources about children? Or maybe it’s my own usage which is unusual in so frequently being about sports and/or children?)

  28. John Turri (#30): I realize (or at least think I do!) that your comment was made mainly in jest. Sorry to respond seriously, but your comment brings up some interesting issues…

    Well, actually, I don’t flat-out assert much of substance in my post. I assert that certain things happened a long time ago (and then in the more recent past), and that when I tried certain searches, certain results were obtained. And I ask some questions. But when it comes to philosophical/linguistic substance, not much by way of flat-out assertion. That ‘know’ is wildly popular is hedged by “at least according to Google.” Even about “know(s)” just being philosopher-speak, I don’t assert that that claim is wrong, but merely say “it looks to me as if that’s wrong” — and I do think I knew that it did look that way to me! (Similarly, I don’t flat-out assert that my interviewer was full of shit, but only that I thought he was. And my claim about the relative fame of Spice Girls & American philosophers, which you dispute (I guess correctly) in #6, is hedged by “I suspect.”) It’s interesting how dangerously substantive flat-out assertions might not be as common as we think, here and elsewhere. This is a point worth keeping in mind when evaluating accounts of assertion (like KAA, which you allude to).

    As for whether I “knew” the substantive things I might have asserted by the epistemic standards that would have governed those assertions….
    I believe that “know” can sometimes be governed by unusually lax standards, at which not much more than true belief is required for knowledge. There’s a section in my forthcoming (supposed to be out May or June ’09) contextualism book on the “floor” for ‘know’ which goes into this a bit. The substantive epistemology that I marry to my contextualism actually invites the *possibility* that ‘know’ sometimes only requires true belief (as a limiting low-standards case). But I just don’t find the needed evidence that ‘knows’ ever goes that low. I consider unusually-low-standards cases that some folks — John Hawthorne, Crispin Sartwell, Saul Kripke (in his “Nozick-bashing” lectures) — have put forward, but while I take them to be claims of knowledge that our theory should accommodate, I don’t think they show that knowledge ever collapses all the way down to mere true belief. Still, I do think the floor does get pretty low, even if it doesn’t get all the way down to mere true belief. And I do think there are allowable standards at which I did count as “knowing” the substantive things I might have asserted. (I won’t go into what more than true belief seems to be needed for “knowledge” by ultra-low standards, but they’re not that demanding at all, and it does seem to be something I had in this case.) And we’ll suppose those substantive things were in fact true. Still, I doubt I knew them by the standards that would have governed my assertions. Though the standards governing assertions in blogs are probably pretty low, I doubt that they’re ultra-low or low enough to count the substantive things I might have asserted as knowledge (for me). Perhaps that’s why I (instinctively, I suppose: I don’t recall consciously wondering whether I knew this or that) hedged on or completely avoided what otherwise might have been flat-out assertions on those substantive matters.

  29. 32: I guess what was written here amounts to agreement with my claim in (29) that googling can be a reasonable place to start.

    On to more substantive matters: The real issue I was trying to address was whether you (Keith) were entitled to draw the conclusion you were drawing about the population based on googling. The point about publishing was meant to draw out the intuition that in order to make a legitimate claim you needed better evidence than just typing ‘knows’ into a search engine (it appears you now have taken-to other forms of evidence which I think is great).

    I apologize if my not agreeing with your original basis for making certain claims about the usage of ‘knows’ was “unhelpful”.

  30. Keith,

    Yes, #30 was mainly jest. And I agree that you never did assert the subsequently hotly disputed claim about Google results. It was clearly qualified — probably more a conjecture than anything.

    I’m glad to hear (appropriately qualified reporting of the) news about your book being out this spring. I look forward to reading it.

  31. On Keith’s #32, there are several things going on. “Knows” takes wh- complements, unlike “thinks” (and under negation you can get “doesn’t know if..” as well as “doesn’t know that..”).

    Plus there are definitely differences in adult vs. child usage. Kids seem to favour “knows” over “thinks” proportionally much more than adults. (See e.g. Frank & Hall, Polysemy and the Acquisition of the Cognitive Internal State Lexicon , Journal of
    Psycholinguistic Research, 20:4, 1991 — but notice that Frank and Hall find “knows” still beating “thinks” among the grownups, whatever Corpus Linguistics might tell us.)

    Anyway, one can’t always figure out what is going on with the mental state ascriptions from looking at the words used, and this is especially true for children. Kids famously don’t really get the idea of *belief* — as a state which may or may not represent the world accurately — until they are almost 4. But they start using “thinks” as well as “knows” in their second year (whatever they mean by it).

  32. What’s in that first paragraph would seem to be huge. (I can’t really speak to the rest, on the children/adult difference.) That the first study would be counting know-wh ascriptions (under the rubric of “mental states ascriptions”), while the other is only counting uses of ‘know’ that control a that-clause would seem to be able to account for a lot.

  33. I know this is a little (one month!) late, but you might want to take a look at the first couple of pages of Wittgenstein’s _On Certainty_ for some insight into what that philosopher probably didn’t, but maybe should have had in mind. Wittgenstein doesn’t say that no one ever says that they do or do not know, but he raises some doubts about whether we can (or should?) use ‘know’ like that. See in particular paragraphs six, ten, and eleven.

Leave a Reply

Your email address will not be published. Required fields are marked *