Micro- Nano-Stylistic Variation
« previous post | next post »
"Don't miss the most loved conference by Delphists like you!"
Philip Taylor wrote to complain about that phrase, which apparently arrived in an email advertisement:
"The most loved conference …" ? I would have written "The conference most loved …".
But his preference apparently disagrees, not only with the author of that flyer, but also with most other writers of English. And it's wonderful how easily we can now check such things. As Yogi Berra (may have) said, "Sometimes you can see a lot just by looking".
In this case, we'll look at relative frequencies of relevant patterns in some of the text datasets indexed at english.corpora.org.
Let's start with (the mere billion words of) COCA:
| Pattern1 | Count | Pattern2 | Count | Ratio |
| the most ADJ NOUN | 114211 | the NOUN most ADJ | 2093 | 54.5 |
| the most loved NOUN | 28 | the NOUN most loved | 3 | 9.3 |
| the most ADJ NOUN by | 387 | the NOUN most ADJ by | 9 | 43 |
Or we can try the same thing on the 15.8 billion words of the NOW dataset:
| Pattern1 | Count | Pattern2 | Count | Ratio |
| the most ADJ NOUN | 1318478 | the NOUN most ADJ | 4780 | 275.8 |
| the most loved NOUN | 3499 | the NOUN most loved | 71 | 49.3 |
| the most ADJ NOUN by | 3174 | the NOUN most ADJ by | 114 | 27.8 |
You're welcome to delve further into the results by country, by context, by time period, by genre, etc. But it's clear that Philip's preference is not a fact about the English language in general, at least not at the level explored by these patterns.
It remains uncertain why (and to what extent) different individuals have strong but idiosyncratic opinions about grammar and style in their native language.
Zanuttini and Horn's edited volume "Micro-Syntactic Variation In North American English" explores the idea that some of this variation is dialectal, correlated with the usual sociolinguistic categories of geography, ethnicity, formality, and so on.
It's possible that Philip's reaction is of that kind, shared with other elderly Brits with classical educations. But it may also be entirely individual.
When we zero in on such minute questions, each person will have encountered a relatively small (and often idiosyncratic) number of relevant examples, interacting in a complex way with the rest of their linguistic, social, and emotional history.
[Of course there are other perhaps-relevant variables in the original example, including the facts that the "most loved" noun phrase is in the object position of an imperative, that the ADJ is a verbal participle, etc. But at that point we're moving beyond the reach of search patterns on words and word-classes applied to merely 15 billion words or so…]
Update — as noted by many commenters, it matters whether the modifier is an adjective or a participle. But there are plenty of examples of the "the most VVD NOUN by" pattern, e.g.
Ferdinand Cesarano said,
September 6, 2022 @ 9:25 am
A note on Yogi's quote: it was "You can observe a lot just by watching."
[(myl) It probably wasn't actually Yogi Berra… But the web finds 3,110,000,000 hits for the "see a lot" version vs. 33,300,000 for the "observe a lot" version.]
Philip Taylor said,
September 6, 2022 @ 9:37 am
To be fair, it was not about "the most loved X" that I was complaining — it was "the most loved X by Y" (in the present case, "the most loved conference by Delphists like you"). It is not "the most loved conference" (of all conferences) — it is rather "the conference most loved by Delphists …".
[(myl) Look at the tables – when "by" is included, you still lose by a factor between about 9-to-1 and 50-to-1 …]
Gregory Kusnick said,
September 6, 2022 @ 10:58 am
To me, "the conference most loved by Delphists like you" says something different and more restrictive than "the most loved conference [organized] by Delphists like you". The first implies that there might be Delphists who (unlike me) love a different conference.
mollymooly said,
September 6, 2022 @ 11:06 am
To me "the most loved conference by Delphists" means something different from "the conference most loved by Delphists". In the second certainly, and the first probably, I would classify "loved" as a part participle or "verb -en" rather than an adjective. What are the numbers for that?
[(myl) I don't have time to run more queries right now, but in all the sequences with by the "ADJ" categories is basically always a past participle anyhow, e.g.:
]
mollymooly said,
September 6, 2022 @ 11:32 am
"in all the sequences with by the "ADJ" categories is basically always a past participle anyhow" I get that for noun most adj but not for most adj noun:
FREQ TOTAL 3,174
THE MOST RELEVANT EXPERIENCE BY 2116
THE MOST TRADED STOCK BY 309
THE MOST ACTIVE STOCK BY 220
THE MOST PIERCING SPEECHES BY 101
THE MOST DISGRACEFUL PERFORMANCES BY 95
THE MOST RECENT REPORT BY 75
THE MOST RECENT DATA BY 59
THE MOST IMPORTANT FACTOR BY 57
THE MOST RECENT SURVEY BY 51
THE MOST RUSHING YARDS BY 46
THE MOST EXPENSIVE WORK BY 45
TOTAL 3174
0.762 seconds
Jonathan Smith said,
September 6, 2022 @ 11:37 am
While the disfavored wording by famously peevish Philip Taylor appears on the interwebs ("Elon Musk is the Most Loved Billionaire by the Internet"; "we will talk about one of the most loved types of hashish by tasters"), I suspect the group that would sniff at this arrangement is rather broad (*cough*)…
Philip Taylor said,
September 6, 2022 @ 11:55 am
Yes, well, if it does mean "the conference organised by Delphists which is most loved" (a possibility that never entered my mind), then "the most loved conference [organised] by Delphists" is unexceptionable, IMHO. But I dispute this interpretation — I believed, and continue to believe, that it is a conference organised for Delphists which (the organisers hope) will be loved by them (the Delphists, that is, not the organisers). In which case I continue to believe that "the conference most loved by Delphists" is the only acceptable wording. As to winning and losing, neither concerns me in the least, only what is right and what is wrong.
Rachael Churchill said,
September 6, 2022 @ 12:37 pm
I agree with mollymooly. I'd like to see the data restricted to -ed adjectives/participles. The "most X noun by" data is swamped by adjectives that can't go after the noun, like "the most recent book by $Author".
Haamu said,
September 6, 2022 @ 2:35 pm
Never mind all that. What's a "Delphist"? Google and Wikipedia (within the limits of my attention, at least) are unavailing.
DaveK said,
September 6, 2022 @ 2:54 pm
“The most loved song by the Beatles” would be the one they recorded that is loved by the most people.
“The song most loved by the Beatles” is the one that John, Paul, George and Ringo agreed was their favorite, whether it was by them or someone else.
JPL said,
September 6, 2022 @ 5:12 pm
I'm not a grammarian (and no authority wrt the terminology), but I thought that in the case of noun phrases, if a modifier with adjectival function has a predicational structure (or, a clausal structure), as opposed to being a single word, then its normal position in the phrase is after the head nominal, which I think is called the "postmodifier" position. E.g., relative clauses; in fact Philip's preferred expression for this example could IINM be described as a "reduced relative clause". Philip's discomfort with the example in the OP is perhaps based on the intuition that the expression breaks up what is in intention a single (complex) modifier into two discontinuous parts, and so expresses what is felt to be the intended meaning less effectively than keeping the modifier expression together after the head nominal.
(BTW, the first clause of Jonathan Smith's comment has the same structure (no doubt on purpose) as the OP example disfavoured by the famously peevish Philip Taylor. That expression does seem more confusing than the one in the previous sentence.)
Terry K. said,
September 6, 2022 @ 6:26 pm
I'm with Philip Taylor in that I would write "the conference most loved by Delphists like you". Assuming the Delphists are being mentioned there as loving the conference, not as creating it. Of course, no reason everyone's idiolect should always match mine.
Also, my inclination is it should be "most-loved" if coming before the noun (in writing). A bit of Googling, though, suggests I'm in the minority there (though "most-loved" is out there), and even some logic why that's wrong. To me, it feels like a unit, a compound, when in front of the noun.
Terry K. said,
September 6, 2022 @ 7:12 pm
I find the example by DavidK interesting. I agree with him on the difference in meaning in that pair. I think, in well edited writing at least, you wouldn't have "the most loved song by the Beatles" with the meaning of a song the Beatles love.
But there's a difference between that example, and examples, such as in Jonathan Smith's comment, where it is clear that the group following the "by" are the ones doing the loving, rather than some other role such as creator.
In the original example with the Delphists, it seems like the Delphists are both the ones who created the conference and the ones who love it, though I would say in the sentence in question "by Delphists" is meant to be the people doing the loving.
J.W. Brewer said,
September 6, 2022 @ 7:24 pm
I think this is the website for the conference in question, which has pretty obviously been written by people who are not L1 Anglophones. http://www.itdevcon.it/en/ My ear agrees that the construction that struck Philip Taylor odd is unidiomatic-sounding, and I thus tend to agree with others that myl's corpus-search results are somehow not operating at the right level of generality to correctly assess how weird-sounding it is. Now, if someone can find a population of L1 Anglophones (as opposed to L1 Italian-speaking computer-programming-conference organizers working in ESL) who consistently produce constructions like this, I'd be interested. I would separately say that someone of Philip Taylor's views ought be able to find plenty of Done-the-Wrong-Way-In-His-Opinion phrasing to condemn without picking on foreigners with an imperfect command of idiomatic English, although maybe the email he got didn't in isolation make the ESL diagnosis obvious enough?
I was not previously familiar with the relevant sense of "Delphist," although given the apparent existence of a programming language and/or software package named "Delphi," "Delphist" as a name for enthusiastic or habitual users of that computer-world thingie seems idiomatic enough. Even though it sounds like a late-nineteenth-century euphemism for a practitioner of some sort of once-taboo sexual behavior now known by a completely different early-twenty-first-century euphemism.
Chester Draws said,
September 7, 2022 @ 12:33 am
In which case I continue to believe that "the conference most loved by Delphists" is the only acceptable wording.
Don't agree to be honest. I hear the intention you seek in the original wording. I can't not hear it, in fact, excepting by forcing another interpretation.
As to winning and losing, neither concerns me in the least, only what is right and what is wrong.
A true peever. Not bothered by what all the rest of the world do. They are clearly wrong.
Bob Ladd said,
September 7, 2022 @ 1:39 am
I agree that Mollymooly's point is important, because, like Jonathan Smith, I find myself in the relatively unusual position of sharing Philip Taylor's feeling that something is wrong with the construction under discussion. J W Brewer may have the explanation. (In any case his gloss on "Delphist" has made my day.)
GH said,
September 7, 2022 @ 3:00 am
I also find myself in the unusual and disconcerting position of agreeing with Philip Taylor. Therefore, I think it is worth taking a closer look at the data. In COCA, the top ten hits for the pattern "the most ADJ NOUN by" (meant to represent the structure we object to) are, with frequencies:
THE MOST RECENT DEVELOPMENTS BY 13
THE MOST RECENT STUDY BY 9
THE MOST RECENT SURVEY BY 7
THE MOST IMPORTANT MEANS BY 6
THE MOST RECENT WORK BY 6
THE MOST RECENT REPORT BY 4
THE MOST RECENT REVIEW BY 4
THE MOST RECENT ATTEMPT BY 3
THE MOST RECENT POLL BY 3
THE MOST STRINGENT TERMS BY 3
It seems to me that these are clearly different in some way from "the most loved conference by…," and that they are not objectionable in the same way. My immediate intuition is that it has to do with "loved by" forming a unit in a way that "recent/important/stringent by" do not.
Terry Hunt said,
September 7, 2022 @ 4:31 am
I'm with Philip Taylor on this one, which may show no more than that our ages and idiolects are similar. For me the phrasing suggests:
* probably not a native British-English speaker;
* possibly a poorly-educated native English speaker;
* possibly not a native speaker of any variety of English;
* possibly an attempted scam.
While of course I could be wrong about all of these impressions, that I formed them does not speak well of the sending organisation's marketing acumen.
GH said,
September 7, 2022 @ 5:46 am
Having examined the corpus results in more detail and experimented with other search variations, I am now convinced that we are dealing with two different structures.
In the more common case (e.g., "the most recent survey by X"), the phrase introduced by "by" refers back to the noun or noun phrase as a whole. In this case, a post-modifier would be grammatically incorrect:
*"the survey most recent by X"
In the less common case (e.g. "the people most surprised by X," "the daiquiris so beloved by Hemingway"), the phrase introduced by "by" refers back to the "adjective" (usually, but not always, a verb in the past tense; cf. "a crime potentially punishable by X"). In this case, reordering the sentence creates a result that is awkward or ungrammatical, or changes the meaning to the first form:
?"the most surprised people by X"
?"the so beloved daiquiris by Hemingway"
?"a potentially punishable crime by X"
The ratios between the two sets of results are therefore not meaningful. You would have to go through the ones that follow the more common pattern to see if the "by"-phrase actually refers to the adjective rather than the noun-phrase. Examining the first 100 hits of "the most ADJ NOUN by" on COCA, only a handful seem like candidates:
THE MOST ADMIRED WOMAN BY
THE MOST DISCUSSED SOFTWARE BY
THE MOST DISAPPOINTED PLAYER BY
THE MOST DEMONIZED GROUP BY
THE MOST DEGRADED HABITAT BY
THE MOST CONSPICUOUS PLANT BY
THE MOST COMPARABLE DEFECTION BY
All of these are false positives (for example, the larger context of the first is "she was named the most admired woman by Americans," so "by Americans" refers back to "named," not "admired") except for one:
(from the blog anarchistnews.org)
Using the search variation "the|a|an ADV ADJ NOUN by NOUN+" I also find only one bona fide instance among the first 100 hits:
(from the Christian Science Monitor)
If this ratio holds throughout, with 1/100 hits actual instances of the construction in question, the version preferred by Philip Taylor, myself, and other commenters is about 2 to 3 times as common as the version we object to.
TonyG said,
September 7, 2022 @ 6:58 am
I shared Philip Taylor's misgivings, so I was surprised by your COCA results. But then after reading GH's comment I realised that we should be searching for past particplies, not adjectives. And If you search COCA for past participles (I registered specifically for this!), the results are even more conclusive, but in the other direction: "the most _vvn NOUN by" returns 2 (two) hits, and "the NOUN most _vvn by" returns 171 hits. Which vindicates Philip Taylor spectacularly!
By the way, I am an elderly Brit with a classical education. I just never realised it until I read your post…
languagehat said,
September 7, 2022 @ 7:29 am
Thank you, TonyG! I was baffled by Mark's post and the general acceptance of the data presented (aside from various forms of nitpicking); it was immediately obvious to me that he was comparing apples to oranges.
Andrew Usher said,
September 7, 2022 @ 8:16 am
I assume that the post and its acceptance was in part due to antipathy toward Philip Taylor; but even if he is just a cranky old peever, he could be right some of the time, and here he made a judgement that is general to native speakers. Linguists ought to be familiar with this kind of grammar – the other posts have given the data, I will just give the simple rule to which it can be summarised:
The 'most' is irrelevant. The rule is that a past participle followed with a by-phrase of its agent gets moved after the noun, while otherwise it takes the default position of adjectives, i.e. before it. This does indeed have something to do with a perception of the two as one unit, and that is due (I am sure) to the analogy with ordinary passives with the same wording, so
"The conference (is/was) [most] loved by Delphists" just needs to lose the copula to become the NP here in correct order.
This is a bit of descriptive grammar not explicitly taught, it seems, but the solution is, as is the usual case, simple.
By the way I'd never seen/heard Delphist in any context before this, if I needed such a word I would have picked Delphian; for the computer language, though, either would seem too slangy to use in writing.
k_over_hbatc at yahoo dot com
TonyG said,
September 7, 2022 @ 3:25 pm
Plenty of examples? Literally all of the examples in your edit are "the most asked question by". This is obviously influenced by FAQ = Frequently asked Questions. And it still leaves a huge preponderance of examples going the other way. Don't you think you should gracefully admit defeat?
Andrew Usher said,
September 7, 2022 @ 5:32 pm
I agree. And the phrase "Frequently Asked Questions" is itself an example of the rule, as effectively no one would choose that order if an agent were expressed: "Questions frequently asked by our readers" would be the full form.
Bill Burns said,
September 7, 2022 @ 6:53 pm
Here's another native English reader and writer who agrees with Philip Taylor that the wording he quoted stopped me in my tracks. The analysis by subsequent commenters I think makes it quite clear that Mark's research was wrongly directed.
Don said,
September 8, 2022 @ 7:03 am
Of course there are “plenty of examples.” If you find the time, I think you would be doing your readers a service by posting a more detailed followup with the correct comparison presented in a similar format to the misleading comparison in the original post.
On a tangential note, Philip Taylor’s comments are one of my favorite parts of LLOG and I don’t know why there would be any antipathy toward him as alluded to by some other commenters. (I do not accuse myl of that antipathy.)