Summary: Things that no languages do
|Submitter Email:||click here to access email|
On May 17th I posted the following query:
''I am interested in collecting examples of phenomena that are not found in any language in the world (as far as we know), where there is no OBVIOUS functional explanation for that fact. Here is an example of the sort of phenomenon that I am looking for: In no language do grammatical processes pay attention to 'third position' (though of course 'second position' is often important).I suspect also that there are many conceivable syntax-phonology and semantics-phonology interactions that are logically possible and not obviously dysfunctional, but which never occur. If anybody has examples of this sort (or, even better, knows if there already exist compilations of them), I would be very grateful to know about them.''
I would like to thank the following people for their replies: Christopher Bader, Ilhan Cagri, Andrew Carstairs-McCarthy, Marina Gorlach, Suzette Haden Elgin, Harald Hammarstrom, Jasper Holmes, Laurence Horn, Angelika Kratzer, Martha McGinnis, Eric Raimy, Uri Tadmor, Jess Tauber, Luis Vincente, Beate Waffenschmidt, Helmut Weiss, and Rok Zaucer
I was reminded that Smith & Tsimpli's book ''The Mind of a Savant'' describes their attempt to teach the savant Christopher an impossible language, Epun, which has some features they believe are not found in any natural language.
I was also reminded that 'third position' (from the end) can be relevant in phonology (though not in syntax).
Here are the suggestions in unedited form. (Some of the properties are not universal, in my opinion, and others have pretty obvious functional explanations, but I repeat them as they were given to me):
There is no attested 'ternary foot' reduplication pattern.
Reduplication is (apparently) never used to mark case although it is used commonly for other inflectional categories as aspect, tense, plurality, etc.
No language favors prosodic palindromes over non-palidromes.
No (known, natural etc) language has developed a place-value system for its numerals. There is a marginal in spoken Samoan where there are place-value variants for 11-99 alongside two other more common expressions (that are the only ones allowed in the written language). I have not found any others in a 700+ sample. However it does not hold for sign languages.
Davis and Koenig (2000: 59-60) claim that there are certain thematic role combinations that do not appear in any language and that there are no linking rules that refer to thematic roles by their relative position in the hierarchy (only rules referring to specific thematic roles). They use both
these considerations to support the view that predicator classes are ontologically more basic than the thematic grids they support.
As ''Contemporary Linguistics'' (O'Grady) claims, no onset of any syllable in any world language consists of 'lp'. There must be some phonological (probably articulatory) reason for that, but I have no information whether it has been researched.
Another thing surfaced here, on Linguist List, and was about writing systems: no language uses bottom-to-top direction, while some use top-to-bottom and many right-to-left or left-to-right.
There is no word for ''not being thirsty anymore'', i.e. the ''liquid'' equivalent of being full. In Germany, there was a campaign tow years ago to invent such a word, but as far as I know the word has never made its way into common usage.
Tenses - is there in any language such a thing as a future conditional?
No languages apparently allow more than four arguments per verb. At least, Pesetsky (1995) claims there are four in examples like this: ''[Mary] bet [Bill] [$10] [that the Calgary Flames would win the Stanley Cup].''
No human language repeats every word twice, as in ''The the bird bird woke woke everybody everybody up up.'' There doesn't seem to be any obvious functional reason why that doesn't happen, unless ''it would be so boring'' can be considered a functional reason; it would be nothing more than an extension of reduplication, which _does_ exist. No human language follows an utterance immediately with its mirror image, as in ''The bird woke everybody up up everybody woke bird the.'' No human language contains a phoneme that requires the lower teeth to be pressed against the upper lip. No human language contains a phoneme that's made by slapping the cheek with the left hand. No human language contains four hundred separate phonemes. No human language has verbs but no nouns, or nouns but no verbs. No human language lacks a mechanism for negation, or a mechanism for number. No human language marks relative clauses by adding a single relativizing morpheme to every word in the clause. No human language uses a beneficiary case marker as a mechanism for marking number, or a number marker as a way for indicating beneficiary case. No human language has a set of morphemes of a particular kind (determiners, for example, or classifiers) for which the choice of the appropriate morpheme depends on the day of week when it is spoken. (For some of these items there are obvious functional reasons that rule them out, but not for all of them.)
I don't think there are any languages that have segmental-phonological conditions on word order (e.g. objects beginning with obstruents precede the verb, otherwise they follow the verb).
Languages do not seem to have the capability to ''count''. It's either none, one or all. So there are languages with no overt Wh-movement, movement of the highest Wh-element, or movement of all Wh-elements. This seems to be true of specifier positions, either a unique one or an unlimited number.
There seems to be no language that has a lexical item for 'not all'. Languages have a word for 'all' or 'none' but not a single word for 'not all'. There do not seem to be words for complements either, as in 'all but three'. Logically, there could be a morpheme for this: 'n'. Thus, 'nthree' would mean all but three, 'nfour' would mean 'all but four'. This could even be productive and
could mean everyone except the children, 'nchildren'. I think this has something to do with entailments; in natural language, entailments must be downward.
It also seems that a structurally case-marked element cannot move into a structural case-assigning position, i.e. an accusative DP cannot move to [Spec, TP] whereas a dative or locative DP can.
I think that a possessor-possessee DP cannot be the external head of a restrictive relative clause: ''John's car which I washed ...'' vs. ''the car which I washed ....''
Implicational universals: Languages cannot have nasal contrasting vowels unless they have oral contrasting vowels. Languages cannot have long contrasting vowels unless they have short contrasting vowels. Languages cannot have fricatives unless they have stops. Languages cannot have voiced obstruents unless they have voiceless obstruents. Languages cannot have affricates unless they have both stops and fricatives. Languages cannot have inflectional affixes unless they have derivational affixes. There no languages where the inflectional affix is closer to the root than the derivational one. Language cannot have prepositions if they have no prefixes. Or: no postspositions if no suffixes.
I don't think there's any language that embeds the independent clause inside the dependent one. And I'm sure no language uses mirror image (''bookkoob'' for ''book'') as a standard word-form.
We have languages like English, in which both nominal and clausal objects appear postverbally (V DP CP). We also have languages like Dutch and German, where nominal objects are preverbal and clausal objects are postverbal (DP V CP). Third, we have languages like Basque where both nominal and sentential objects are preverbal (DP CP V, though there is also an option for a Dutch pattern DP V CP). However, the fourth expected option, which we might label ''mirror-Dutch'', does not exist, as far as I know. This language would feature nominal objects obligatorily in post-verbal position, and sentential objects obligatorily in preverbal position (CP V DP).
There is a short squib by Mark Baker in an LI issue from a couple of years ago, where he discusses the interaction of parameters that regulate V-to-T movement subject raising to SpecTP. So, we have languages in which both the subject and the verb raise to the TP area, such as French. We also have languages in which only the subject raises to TP, while the verb stays in VP, such as English. The third option are languages, such as Irish, where only the verb raises to the TP area, and the subject stays inside VP. However, the fourth expected type, a language in which both verb and subject stay inside VP, is not attested. This language would have a SVO order in simple tenses, and an Aux SVO order in compound tenses.
No language so far as I know lacks onomatopoeia of some sort. Also no language may lack at least a small handful of ideophones, which relates to the first point since the latter tend to also be onomatopoeic in languages with small numbers. In languages with large numbers, the ideophone system tends to branch out into other areas of sensorimotor reality,
Chapter 2 of Carstairs-McCarthy's book ''The Origins of Complex Language'' was about things no languages do. Somewhere else in that book it is mentioned that no language places phrasal modifiers on whichever side of the head is closer to the verb, as in:(1) children small many played beach on 'Many small children were playing on the beach' (2) on beach played many small children 'Many small children were playing on the beach'.
|Original Query:||Read original query|
Sums main page