Occasional commenter Martin writes to ask about adjectives:
The rule in English apparently goes something like, “number + judgement/attitude + size + age + color + origin + material + purpose + noun.” (Wikipedia has an expanded and annotated list.) But there are exceptions, for example “modifying adjectives that are homophonous with reduced relatives, or exhibit a special intonation pattern (such as ‘comma’ or focus intonation) are allowed to escape ordering restrictions.” That’s from this 2006 paper by Alexandra Teodorescu which is focused primarily on this and other exceptions.
One question around the ordering rules is, why are they that way? Why is it that “the strange old Polish ladies” sounds correct to our ears, while “the Polish old strange ladies” does not? This 2017 paper by Gregory Scontras, Judith Degen, and Noah D. Goodman argues that subjectivity governs the standard ordering, with the most subjective adjectives being placed the farthest away from the noun being modified. (It would seem that quantity or number is an exception to this rule, which the authors don’t mention in the paper.)
But in their final discussion the authors acknowledge that their findings about subjectivity just raise another “why” question: “While subjectivity accounts for the regularities we observe in adjective ordering, the deeper explanation for how subjectivity determines the relative order of adjectives remains unsettled.” They continue:
For now we can only speculate about the ultimate source of this desire. Subjective content allows for miscommunication to arise if speakers and listeners arrive at different judgments about a property description. Hence, less subjective content is more useful at communicating about the world. An explanation along these lines, based on pressures to facilitate successful reference resolution, would have to depend on the hierarchical, not linear, ordering of adjectives: noun phrases are built semantically outward from the noun, and more useful, less subjective content enters earlier in this process (cf. the mirroring of preferences in pre- vs. postnominal languages). A full explanation must examine not only why we observe the preferences that we do, but also how and to what extent these preferences get conventionalized via the diachronic processes that shape language—a promising direction for future research.
Whatever its source, the success of subjectivity in predicting adjective ordering preferences provides a compelling case where linguistic universals, the regularities we observe in adjective ordering, emerge from cognitive universals, the subjectivity of the properties that the adjectives name.
This conjecture does start to explain why this particular grammar rule feels so natural or internalized, and maybe is less subject to gradual change over long periods of time, as compared to many others which feel (and are) more artificial and likely change, for example not ending sentences with prepositions or splitting infinitives.
Then there is the question of how the ordering rules vary among different languages, which I can’t find much about. Apparently most languages have rules for the order of adjectives, but does the subjectivity rule apply generally in other languages?
Thoughts?
I was taught part of the rules for English explicitly in school; in German they’re largely similar, but on the whole less strict. I was not taught the subjectivity business, and right now I’m too tired to think about it much. 🙂
Green colorless ideas sleep even more furiously…
There is a paper by Richard Sproat and Chilin Shih, The Cross-Linguistic Distribution of Adjective Ordering Restrictions (in Interdisciplinary approaches to language : essays in honor of S.-Y. Kuroda, 1990, p. 565) which I never got all the way through. I see on Google Scholar that it’s been cited some 500 times since.
I went to the paper to figure out what adjectives homophonous with reduced relatives are, but I think she made a typo in the mandarin set there, unless the changing spelling is related to her point?
hao-DE yaun-DE panzi
yuan-DE hao-DE panzi
Is the homophony she’s pointing to the fact that hao and yau sound similar? (Do they sound similar?) And then it gets lost when she misspells it yuan in the second example? Or is the use of the particle de the thing providing homophony, and yaun/yuan is changing because of some word-form issue?
Neither yau nor yaun are possible syllables in Pinyin transcription; yaun is definitely a typo for yuan.
It’s yúan everywhere I can find. I don’t see where the typo is.
For Chinese, they argue that adjectives compounded with 的 de can occur in any order, but without it they follow constraints similar to English. I don’t follow the theoretical semantic framework they use, but I think they are saying that adjective-的 compounds function like relative clauses. In English, too, relative clauses are not constrained in their order.
“ This conjecture does start to explain why this particular grammar rule feels so natural or internalized, and maybe is less subject to gradual change over long periods of time, as compared to many others which feel (and are) more artificial and likely change, for example not ending sentences with prepositions or splitting infinitives. “
As an aside, not ending sentences with prepositions or (the proscription against) splitting infinitives are not rules, but stylistic preferences. The violations of such guidance began centuries before anyone attempted to enforce such “rules”, and mercifully have continued to the present day.
Indeed, to build on cuchuflete’s point, one interesting thing about the adjective-ordering “rule” (or strong default pattern) is that it is AFAIK not usually ever explicitly taught in school to L1 Anglophone students in Anglophone countries (as opposed to ESL students), yet people “obey” it anyway. Amazing! It’s as if native speakers of a language tacitly understand its grammar w/o schoolteachers having told them what it is. Who woulda thunk it?
That said, I think the “expletive infixation” rules are an even stronger example, since L1 Anglophones typically have quite uniform opinions as to where you can and can’t infix expletives and we can be even more confident that essentially none of them were ever taught those rules in school. (I was “taught” those “rules” in school as a 21-year-old linguistics major taking a morphology class, but I already tacitly “knew” and followed them.)
Kusaal has the Oti-Volta thing whereby adjectives are compounded with a preceding noun stem, as with bʋʋg “goat”, bʋpielig “white goat.” It is possible to get informants to add another adjective, e.g. ninwɔk.pielig “white tall person”, but examples in the wild are hard to find, so the question of ordering hardly arises: it doesn’t seem to be a very natural construction regardless of order. (Adjectives, do, however, quite freely form non-final components of compounds, because dependent pronouns like demonstratives and indefinites also compound with preceding stems, e.g. bʋpielkaŋa “this white goat.”)
Quite a lot of what would be adjectival meanings in English etc are expressed by using a free noun as a premodifier: so, for example, with expressions describing what something is made of: salima bʋtiŋ ” golden cup” (salima “gold”); or expressing ethnicity, e.g. Nasaal bugum “electricity” (“European fire”, where Nasaal alone is actually “European language, English/French.”)
Rather surprisingly, these uncompounded premodifiers bind tighter semantically than following elements of compounds do: [[salima bʋtin]tita’a]kaŋa “this [big [gold cup]]”; even more surprisingly, this is still true when the premodifier consists of more than one actual word:
[[anzurifa nɛ salima] la’]maan
“[[silver and gold] goods]-maker” i.e. maker of silver and gold items.
The typo yaun is on p. 2 in example 7.
What is homophonous then? Just the two particles? Are they homophones or just repetitions of the same particle.
Ryan, I think we’re looking at different places. I’m looking at the paper in the Kuroda book I quoted above. There’s no p. 2 or example 7.
Sorry. My initial comment came after yours but was in reply to the main post. The homophonous adjectives quote is from the linked 2006 paper. Or is that the 2006 linked paper?
Example 7 on p2 of the 2006 paper quotes from the 1991 paper by Sproat and Shih. In the 2006 paper, there is a misspelling of “yaun” for “yuan” in the quote from Sproat and Shih.
@Ryan, the homophony, as I understand it, is that, for example, “hao de x/好的X” (“good x”) is homophonous with a relative clause with a similar meaning (“x that is good”) in Mandarin. For a contrived example, if one heard “hao de che 好的车” it could be “the good car” and “the car that is good”.
In general: https://wals.info/feature/87A#2/18.0/152.9 (but there seems to be nothing about the order of adjectives within a noun phrase).
In French noun phrases:
– Adjectives usually come after the head noun but sometimes before, and a specific adjective is most often in a fixed position relative to the head noun;
– Some adjectives can go on either side, but they usually have a different denotation or connotation in each order;
– Postnominal adjectives are separated by “et” (“and”);
– Overwhelmingly no more than one prenominal adjective, and if more than one they’re not separated by “et”;
– I need to think about the order of postnominal adjectives (after I get some sleep), but I believe relevance or centrality to the topic plays a role.
There was apparently some manipulation of this in the “Forrest Gump” movie, where he says “a blue clear sky”. (I didn’t see the movie.) Then later there was a popular song written with the same title, following the movie.
It doesn’t sound right to me, but as a native speaker, I never explicitly learned these rules. “Wild blue yonder”, “great white North”. I think that colour terms bind more closely than (some) others.
If I were teaching ESL students, I think I would tell them, forget learning the rule, just don’t try to say things like “the worn-out old green French knitted woolen fisherman’s hat”. Just use your adjectives one at a time.
In German sentence order we get taught STOMP–subject, time, object*, manner, place (*indirect or direct, order depends on if they are pronouns). Although there seem to be exceptions. But I’ve never seen anything similar for adjectives. Maybe the languages are similar enough that we English-speakers can just follow our instincts and not sound ridiculous.
It seems to me, in my experience, which is much more limited than many people here, that there are some languages in which authors don’t feel the need to create these adjective piles. It is just not the way people talk.If you tried to translate it into that language, you would have to figure out another way of expressing it.
To me it seems rather Germanic. But I have forgotten most of the Greek I studied a half-century ago. Does Homer ever say anything like “the perilous wine-dark salty-tasting roaring man-killing deadly jellyfish-nurturing foaming wave-filled wind-driven sea”?
The placement of time and place is definitely a lot looser than in English. Learning how strict it is in English was pretty hard.
I certainly hope there’s a jellyfish-nurturing sea in Homer!
Near where I live is a driving school that announces itself as École de Conduite française. Why would anyone want to learn French driving, I thought when I first saw it. However, that was more than 30 years ago, and at that time I found (as a pedestrian) French driving wilder than English. However, in the interim period French driving has improved a lot and English driving has deteriorated a lot, to the extent that I now feel safer crossing the street in Marseilles than in Birmingham. Anyway, I still find the placing of the adjective surprising, and I would expect École française de Conduite.
In example (12) the operator adjective former is present and the word order is free.
Бывший лучший, но опальный стрелок…
Korandje too tends to avoid multiple adjectives modifying the same noun; it can be elicited, but even then often with some difficulty.
As David probably already knows, Nasaal is cognate to Nazarene, via Arabic naṣārā “Christians” and, I dunno, probably whatever the Hausa version of that is.
Sproat and Shih claim that in French, chien moyen blanc and chien blanc moyen are equally valid, as are maison blanche carée / maison carée blanche and piano antique noir / piano noir antique.
@Lameen:
In standard (i.e. Kano) Hausa, Nasara still means “Christians”, as you’d expect. but in western Hausa dialects like that of Ader* (and in Ghanaian Hausa) it’s shifted to “Europeans.” All the WOV languages seem to have adopted it from there.
We had a discussion about West African words for “European” not long ago.
* There’s a nice grammar of it by Bernard Caron (as you probably know.) Students of Hausa really are spoilt when it comes to language description and documentation …
Talking of Hausa, Hausa adjectives can precede or follow the noun; preceding is commoner, and some speakers apparently feel that following adjectives are more emphatic, while others say that there is no actual difference. More research needed …
I don’t think constructions with more than one modifying adjective are common in Hausa; at any rate, I can’t find anything about ordering constraints.
@maidhc, David Marjanović: I first learned the German order for predicate modifiers as: Zeit, Art, Platz. However, I looked at some older and more comprehensive* German language textbooks stated that the standard order was: time, place, manner. My high school German teacher, Mr. Chapman, told me that that was the “traditional” order, although it had been outdated for decades at least.
Whichever order one followed, it seemed confusing to us, as native English-speaking students, that there should be a required order at all. We had no idea that there was anything similar in English. “I am riding my bike tomorrow to school,” is clearly not quite right, but none of us had ever noticed it.
* Textbooks for under-eighteen children differ from those aimed at the college and higher market, in that they have teachers’ editions and sometimes whole packages of supplementary worksheets, filmstrips,** and other materials. One consequence of this is that there is lots of material one is supposed to learn that is not contained in the student texts. This becomes less significant in books aimed at older students. High school algebra books state everything you need to know, with the teachers’ editions just adding recommendations for emphasis and explanations. However, all the high school language textbooks I used left significant amounts of vocabulary and grammar out, and with a German teacher as idiosyncratic as Mr. Chapman, I never felt like I got a full picture of what we were supposed to have learned at each stage. The older textbooks I found, some of which had been my mother’s, were immeasurably better in this regard, stating explicitly and fully everything to be learned in each unit.
** I imagine filmstrips have been largely superseded in the fully digital age, where every classroom has a smart board, but you get the idea
every classroom has a smart board
And thus I learn that there is such a thing as a “smart board.” Alexa, spell “cat”!
I’m sorry, Steve, I’m afraid I can’t do that.
I first learned the German order for predicate modifiers as: Zeit, Art, Platz
There is a weird (weird to me, anyway) constraint in Kusaal that whereas any kind of adjunct can be placed before the clause subject by explicit preposing mechanisms [using ka, which sort-of means “and”, though it can mean all sorts of other things in practice], only phrases expressing time or circumstance can precede the subject without any formal mechanism to prepose them; you can’t do it with expressions of manner or place.
So you can say
Mam bɛ nɛ mɔɔgin zina. “I’m in the bush today.”
Zina mam bɛ nɛ mɔɔgin. “Today I’m in the bush.”
Zina ka mam bɛ mɔɔgin. “Today I’m in the bush.” (with explicit preposing)
Mam bɛ mɔɔgin. “I’m in the bush.”
Mɔɔgin ka mam bɛ. “I’m in the bush.” (explicit preposing)
but not
*Mɔɔgin mam bɛ.
“Whichever order one followed, it seemed confusing to us, as native English-speaking students, that there should be a required order at all. We had no idea that there was anything similar in English. “I am riding my bike tomorrow to school,” is clearly not quite right, but none of us had ever noticed it.”
As a Russian speaker I felt similarly about English adjectives. I still do not understand to what extent it is a “rule”.
Same for us from the other side.
Now that I think about it:
Morgen fahre ich mit dem Fahrrad in die Schule. – emphasis on “tomorrow”, though that can be just to introduce tomorrow as the topic
Ich fahre morgen mit dem Fahrrad in die Schule. – emphasis on not walking, unless intonation overrides that
?Ich fahre mit dem Fahrrad morgen in die Schule. – emphasis on “tomorrow”, but definitely weird
*Ich fahre mit dem Fahrrad in die Schule morgen. – wrong.
Platz isn’t right, BTW; that should be Ort.
@David Marjanović: We were supposed to remember Zeit, Art, Platz using the acronym “ZAP.” That’s a better mnemonic than “ZAO.”
@DM: your “weird” example actually sounds plain wrong to me.
I can sort of make it work with the right intonation in the right context, like talking to someone who’s hard of hearing. But not well. Should have added another question mark or two.
As dm said, the normal way of emphasising morgen is to put it first, morgen fahre ich… Schule. Also what about morgen fahre mit dem Fahrrad in die Schule ich or morgen fahre in die Schule ich mit dem Fahrrad? These seem to me wrong but correct in some close parallel universe. In the second one maybe the bike has a puncture and you have to take it on the tram.
The great distance between the (basically clitic) pronoun and its verb is ungrammatical. It’s very hard to put anything between those two words (in any of the three basic word orders).
Morgen fahre ich in die Schule – mit dem Fahrrad only seems to work with a pause.
Interesting. To me it sounds neither wrong nor weird, just a bit unusual and rare.
It is how it works in Russian. I mean: you find an appropriate context, intonation and structure that would motivate your “wrong” word order.
When someone asks you if it (someone’s word order) is “natural” without this, it may sound wrong.
I also assume, when you ask someone if somethign sounds natural, and she says “no”, the scientific fact that you learned is, precisely: “when asked out of the blue, whether this line sounds natural, some native speakers say no”.
Where are you from?
I found another way to make it work: by taking the bike as the topic from earlier in the conversation. Ich fahre mit dem Fahrrad morgen in die Schule, übermorgen einkaufen…
I did not find the way to make my Russian example that did not work work. I just slept and when I remembered it, it worked:)
It was “я поеду на велике завтра в школу”.
ًWhat worked was dealing with поеду на велике (will-go on bycicle) as a single word/unit. As when you say “I will do it (ride my bicycle)”. Possible if you previously (not necessarily the same day) have been discussing it, using bicycles for transportation. Then within the information structure it is a unit that can be topicalized-focused as a whole, and the line is not differnet from я поеду завтра в школу.
What prevented me from seeing it is that I was putting differnet words in focus/topic by one, treating “on bicycle” as a unit, and forgot that there is this thing, when you deal with a block as a block. It is a problem with any search algorythm or an attempt to prove awkwardness by exhaustion: you move within a finite subset (defined by your model) of the huge space of possible meanings, and you do not notice what your model does not predict. That and just the fact that when you are thinking hard about awkwardness of something, your perception of awkward is terribly skewed.
Having this said, my example is not “awkward” anyway, it is normal in speech – think, what transcribed speech looks like usually – just a bit clumsy. I was trying to make it well motivated, not clumsy, and as I was thinking and my perception was getting skewed it was drifting to odd.
The real terror is that when you are reading Tosltoy, everything is natural, because he is Tolstoy, he knows Russian better than you – at least Tolstoy’s Russian.
And when a learner asks you – everything is awkward, including things you say yourself. At least for some people. I have seen all sorts of odd “ungrammaticality” judgements from native Russian speakers in this situation.
I am a victim of this parasitic ungrammaticality myself even though I know about the effect;(
There was a bit in one of the pop-linguistics books I’ve read (forgot which one) where it started with a four-word Russian sentence, methodically listed all 24 possible word orders, and for the unnatural-sounding ones, gave a context where that particular order would make sense.
Meanwhile, in my actual linguistics classes at university, we were taught (among other things) the work of Guglielmo Cinque on the ordering of adverbs, as an example of the frontier of research in linguistics. Reportedly the ordering is relatively rigid, even across different languages, despite sentences with many consecutive adverbs in them being very rare.
(There were several further works mentioned that built up on it, but I don’t recall the details; to be honest I mostly only remembered Cinque because of his surname.)
It seems in Bavarian (maybe more generally?) you can also put adverbs after the verb in a relative clause for emphasis:
Hochwiern här geischlinger Rad, jäz mus ich es ienen schreim, das mir Ale im Kasino fersambelt gewesd sind am Abend dafor
[ich] sag es deiner Alden mit was fier einen Schlidden Du schbazieren farst in der Schtadt.
Thoma, Briefwechsel eines bayrischen Landtagsabgeordneten
…That’s an attempt to write Standard German by an apparently fictional character who only has a middling knowledge of it and has no clue of the orthography.
Putting times and places sentence-finally like this (am Abend davor, in der Stadt) happens a lot in spoken language, my own included, but should be taken literally as people adding stuff because they notice they need to add more information after they’ve already finished the sentence. When people successfully plan the entire sentence before they say it, this happens much less often, if at all.
The real terror is that when you are reading Tosltoy, everything is natural, because he is Tolstoy, he knows Russian better than you – at least Tolstoy’s Russian.
And when a learner asks you – everything is awkward, including things you say yourself.
I hadn’t thought of that, but it makes sense. Context is all.
despite sentences with many consecutive adverbs in them being very rare
Even two is rather difficult in English; the phrase “meaninglessly Swedishly-named sprockets” for the various objects that Ikea sells has stuck in my mind since 2004 because of its oddity.
Sometimes it happens (silently invisibly).
Медленно неслышно крался… (крался неслышно И медленно)
Неожиданно тихо прошептал (прошептал [неожиданно [тихо]])
Потрясающе красиво нарисован(ный) (participle, used predicatively(attributively))
Неожиданно густо-красный.(adjective, [неожиданно[густо]]красный)
More or less normal.
Adverbs are very easily piled up in German, probably because they’re identical to predicative adjectives.
silently invisibly
That’s different: it means ‘both silently and invisibly’, and I’d write it with a comma. But in the first case, it’s not that 0) the sprockets are meaningless, nor that (1) that the names are meaningless, but that (2) the Swedishly-named nature of them is meaningless.
That’s different
That’s true, but if each adverb is modifying the following one, then the order is determined by the meaning, not any sort of arbitrary rule. Investigating the question of order requires a string of adverbs that are all modifying the same word or phrase, as in drasvi’s example.
GloWbE has hits for things like “only really socially closely”, but I think those fall afoul of Keith’s requirement.
I’ve been able to find a comment calling someone “really ugly politically culturally religiously and primitive”, but then I’m not sure how would one rank those for subjectivity.
I’d say the order is that more permanent attributes come closer to the noun. If there’s a set phrase of attribute+noun (chocolate cake) you can’t place anything in between. Possessive attributes and demonstratives come the farthest since things easily change owners… so in English they developed into articles/determiners: my favorite chocolate cake. Your old Polish friend. An old Russian guy. A red-haired Russian guy. A Soviet Red Guard.
But what is considered “permanent” has some variation across languages, and that is interesting to investigate…
I am not sure about this. Maybe I should have a slice of chocolate sponge cake…