10

I was interested in the following sentence which appeared in an article titled “Colorless, Tasteless but Not Dangerous" by Dwight Garner in The New York Times (November 15, 2010).

People who do gravitate toward these sort of things, he warns, sotto voce, might be “the wrong kind of white person.”

Can someone clarify if the fragment "these sort of things" is ungrammatical, as I think it is?

I would reword "sort" with "sorts", but I'm not sure on this correction because the phrase "these sort of things" occurs on many occasions on The New York Times, it frequently occurs in others newspapers and, more generally, it has 2,670,000 hits on Google Search. So I am wondering if it is in common usage, albeit it isn't the highest register.

  • 1
  • 2
    Except that it's not an exact duplicate. This question deals with "these kind of X", with a plural determinant and a singular word from kind/type/sort, and those questions deal with "this kind" or "these kinds". – Peter Shor Jun 8 '12 at 15:55
  • 2
    This is decidedly not a duplicate of the question it’s been marked as a duplicate of. It’s a completely different question, and none of the answers in the dupe (or the related questions in Matt’s comment) address this question. – Janus Bahs Jacquet Feb 4 '17 at 10:59
  • 1
    This question is nearly 5 years old. I imagine Google's algorithms have changed quite a lot in that time period. – Martin Smith Feb 18 '17 at 19:39
  • 2
    @Martin Smith: The error with the search estimate is a very old one, I remember encountering it years ago when only starting with English and Japanese. When trying to estimate the use frequency with Google, one must ALWAYS go to the next page and then set the start parameter in the search URL to something absurdly high, that will show you the real number of pages and hits. – undercat Feb 18 '17 at 20:16
8
+50

To gauge the relative popularity of "these sort of things" in historical usage, I ran Elephind newspaper database searches for that phrase and for fifteen related expressions, representing various permutations of this/these, sort/sorts, and thing/things across the period 1800–2017.


Interruption: An important note on Elephind searches

Before discussing the results, I should observe three things about Elephind search results.

First, the results tend to tend to yield results whose numbers distribute across a bell curve, owing to a couple of factors: (1) the relative scarcity of newspapers in the databases that Elephind searches as you go back in time; and (2) the consequences of copyright restrictions on the availability of newspapers for searches as you come nearer the present. The net result of these contrary forces is that the number of newspapers searched reaches its maximum in the decade 1910–1919, and in fact peaks in 1915. (I base this conclusion on the results of an Elephind search I ran for the word the.)

Second, Elephind searches substantially more newspapers from Australia than from the United States over the full range of its coverage (1780–2017). My search for the, for example, yielded matches from 104,628,361 Australian newspapers and 42,601,972 U.S. newspapers. That's a lot of newspapers from each country, but it's a lot more from Australia. This reflects the greater accessibility of the digitized newspaper collection through Trove (the online presence of the National Library of Australia) than of the combined Library of Congress and state online newspaper collections in the United States.

Third, newspapers from Australia form an even greater majority of the searchable newspapers from 1925 forward: 47.3 million from Australia to 10.3 million from the United States, judging from my the search. That's because Australian newspapers continue to be available from Trove after 1925, where as the Library of Congress's collection from 1925 forward is not searchable at all. The main sources for U.S. newspapers from the post-1924 era are the state collections from California (3.6 million newspapers), Illinois (2.6 million), Colorado (1.0 million), and Texas (0.79 million).

So the results of an Elephind search give outsize prominence to results from Australia and from states that have searchable newspaper databases (in addition to the states already named, Pennsylvania and Virginia have searchable collections of more than 100,000 newspapers from 1925 onward; and university libraries in Missouri, New York, Ohio, Massachusetts, and Virginia do, too), and they become progressively weaker as you get closer to the present.

Having clarified what the source of my results is, I return to those results.


How usage is distributed in the Elephind results

Here is the distribution of the matches I got for the sixteen phrases that I searched for across the years 1800–2017 (with the source of the matches further identified in parentheses by country of publication):

this sort of thing: 90,278 newspapers (Australia 60,017, United States 30,261)

thing of this sort: 1,834 newspapers (Australia 1027, United States 807)

things of this sort: 1,557 newspapers (Australia 641, United States 916)

these sort of things: 1,437 newspapers (Australia 1,211, United States 226)

this sort of things: 331 newspapers (Australia 192, United States 139)

these sorts of things: 203 newspapers (Australia 142, United States 61)

these sort of thing: 23 newspapers (Australia 17, United States 6)

these sorts of thing: 1 newspaper (Australia 1, United States 0)

things of these sorts: 1 newspaper (Australia 1, United States 0)

this sorts of thing: 1 newspaper (Australia 1, United States 0)

this sorts of things: 1 newspaper (Australia 0, United States 1)

thing of these sort: none

thing of these sorts: none

things of these sort: none

thing of this sorts: none

things of this sorts: none

Clearly the dominant phrasing is "thing of this sort" at 90,278 matches, a figure that dwarfs two forms that I have heard used quite often and that (I believe) no one would be inclined to challenge as ungrammatical: "thing of this sort" (1,834) newspapers and "things of this sort" (1,557).

But then—to my surprise—comes the wording that the poster asks about—"these sort of things," at 1,437 matches. That number is not tremendously smaller than the numbers for "thing of this sort" and "things of this sort," and it leads me to believe that it is a (or was) a fairly common form of the expression in at least some part of the English-speaking world.

The most noticeable thing about the expression is how strongly the matches skew toward Australian sources: the ratio of Australian to U.S. sources is more than 5 to 1, as against slightly less than 3 to 1 for "thing of this sort," a bit more than 2 to 1 for "thing of this sort," and less than 1 to 1 for "things of this sort."

This last ratio is especially surprising, since one might expect that "this sort of things" might make a natural pairing with "things of this sort." But U.S. usage seems to accept "things of this sort" far more readily than it does "this sort of things."


Conclusion: Is 'these sort of things' grammatical?

If the test of grammaticality is usage, the Elephind search results tabulated above offer "these sort of things" some fairly impressive credentials for claiming legitimacy—at least in Australia, where, by a very thin margin, it is the second most common form in historical newspaper database occurrences after "this sort of thing." In the Elephind results from the United States, however, it finishes a less impressive fourth, and out-tallies the fifth-place "this sort of things" by a count of only 226 to 139.

As a practical matter, "this sort of thing" is the runaway winner in historical newspaper use in Australia and the United States. Both "thing of this sort" and "things of this sort" seem to pass the no-one-bats-an-eye test as well (at least, I've never heard anything against them), and that may be true of "these sort of things" in Australian usage, too. But in the United States that last wording is infrequent enough to prompt some degree of eye batting and perhaps even eyebrow raising when encountered unexpectedly.

  • 1
    +1 I've noticed quite a few typos in the Washington Post in the last year or so -- not typos that can be blamed directly on Spellchecker -- but typos that mean the paper is saving $$ on proof-reading, so an indirect result of Spellchecker. I can't remember an example, but they are irritating. – ab2 Feb 19 '17 at 4:09
14

I'm going to go out on a limb and say that the grammar norms suggest that the correct usage would be sorts. However, after looking into the issue more, I was not so convinced. http://www.phras.in had a high instance rate for "These sort of things..." and even Swan's Practical English Usage had an interesting passage about sort of:

When we are talking about one sort of thing, we can use sort of, kind of or type of followed by a singular noun.

This sort of car is enormously expensive to run.

Plural demonstratives (these and those) can also be used.

These sort of cars are enormously expensive to run. Do you smoke those kind of cigarettes?

This structure is often felt to be incorrect, and is usually avoided in a formal style. This can be done by using a singular noun (see above), by using plural sorts/ kinds/ types, or by using the structure ... of this/that sort/kind/type.

This sort of car is ... These kinds of car(s) are. . . Cars of that type are ...

Given this, I would say that at least according to Swan and the current usage on the internet at least, both are correct.

6

Both 'this sort of thing is... ' and 'these sorts of things are...' in use, the latter being less common and informal.

From Michael Swan's Practical English Usage:

enter image description here

enter image description here

Also from The New Fowler's Modern English Usage:

enter image description here

5

As you suspected, the statement is incorrect. These sorts of things is correct, because these is used to refer to plural nouns. This sort of thing is also correct; it is singular. However, These sort of thing has these being used to refer to the singular sort, and so it is wrong.

  • When you get 2.5 million hits on Google, I would posit that brandishing a particular usage as "incorrect" is overly simplistic. – Neil Coffey Jun 8 '12 at 1:21
  • 1
    @Neil Coffrey: Slang is commonly accepted as part of the language, but it's still "incorrect" in a grammatical sense. If a grammar is a set of rules it really gets to be a pain if you need to incorporate every slang phrase as an exception to the rule! – M. Dudley Jun 8 '12 at 1:26
  • 3
    It isn't incorrect because it violates a grammatical rule, as exceptions to those are common. Its incorrect because it still sounds awkward to the modern ear. If it stops sounding awkward and wrong, then it stops being awkward and wrong. – Lawton Jun 8 '12 at 13:53
  • 1
    Except that it's still in use in various regions (mostly in the U.K., I believe) and to people from those regions it sounds perfectly natural. – Peter Shor Jun 8 '12 at 15:53
  • 2
    @NeilCoffey - YOU DO NOT GET 2.5 MILLION HITS ON GOOGLE. – Hot Licks Feb 18 '17 at 19:04
0

The argument that you get 2.5 million hits on Google, or some other enormous number, is wrong. The expression is really quite rare.

When I Google with these sort of things, with no quotes around the argument, I get an estimated 230 million hits. But if I "Next" to the end of the list of "unique" hits, I get only 234 hits. Doing "repeat search with omitted results" it estimates 307 million hits, but lists only 322 if you "Next" to the end. Keep in mind that this list of hits includes web pages that contain the four words in any order and interspersed with other words, not contiguous in the order specified.

When I Google with "these sort of things" (quotes included), I get an estimated 396,000 hits. "Nexting" to the end, Google lists only 249 hits. If I "repeat search with the omitted results included" I still get an estimated 396,000 hits, but "Next" gives up on page 63 (ie, about 640 hits) without claiming to be at the end.

Googling "these sorts of things" (note the plural), I get an estimated 477,000 hits, and "Nexting" to the end it gives up with 255 results. Doing "repeat search with omitted results", it still estimates 477,000 hits, then refuses to search any farther after "Next" counts 300 results.

If I Google for the expression "the first thing that", Google estimates 35,500,000 hits and, with "omitted results", actually lists 324.

It's hard to interpret these results, other than to say that:

  1. Google "estimates" are probably "over-estimates", while the actual number of entries returned is almost certainly an undercount, with a number in the neighborhood of 200-300 being likely. (I don't know if the one outlier has any significance or is just a random oddity.)
  2. If you don't include quotes around a phrase when you Google it, the numbers are entirely meaningless, but if you include the quotes then the "estimates" probably can be roughly compared to estimate frequency.
  • And I'll note that if you look at Ngram you will see that "sort" was oddly more popular than sorts" until about 1970. – Hot Licks Feb 18 '17 at 22:15
0

No, according to The Grammar of English Grammars, 10th edition (1869)

On pages 544-545, under the title "Improprieties For Correction: False Syntax Under Rule IX: Examples under note I.-Agreement of Adjectives" there is the example:

"For these sort of things are usually join'd to the most noted fortune" Bacon's Essays page 101.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.