martedì 28 giugno 2011

Universals of Grammar with Particular Reference to the Order of Meaningful Elements

This paper aims at providing a description of Greenberg’s article Some Universals of Grammar with Particular Reference to the Order of Meaningful Elements.
The paper is structured as follows:
- definition of linguistic typology
- role of universals of language
- Greenberg’s 45 universals and their examination

Towards a definition of linguistic typology
Linguistic typology deals with interlinguistic variation study labelling historical-natural languages on the basis of similarities and divergences in their structure. What this approach underlined was that these variations were not casual but they obeyed general principles called language universals. On the basis of common structural traits, we can classify languages not only from a genealogical point of view but also from a new perspective which takes into account structural organization and resemblances in languages. Hence a linguistic typology can be defined as a whole of structural features in concord among them, which means loosely speaking a group of languages which share many features despite their different origin and geographical position. It should be stressed that typology does not act as a theory of language. It does not aim at providing a formal model of language structure. Rather typology is an approach based on probability and statistics that linguists can employ in order to investigate the composition of human languages thanks to a crosslinguistic comparison which rests on the assumption that structural similarities disclose fundamental properties.

The role of universals
As linguistic and cognitive phenomena, universals have a long tradition in the philosophy of language. Language is here often meant as innately specified in every human being and language universals are then equated with those features of language that are part of man’s genetic endowment. Hence, linguistic typology and universals are strictly related, because the latter are recurring properties in the language structure regardless of its genetic relationship and potential reciprocal conditioning. The universals of language which must be meant as universal tendencies can appear in the form of:

- implicational universals, of the sort of: if Language B has feature x, then it will tend to have feature y
- nonimplicational universals, i.e. all languages tend to have feature z.

Greenberg and many others operating in the framework of what came to be known as the typological approach found out that an analysis of a substantial number of languages revealed not only the range of variation but also constrainted on that variation showing that languages do not vary casually.



Greenberg’s universals
Greenberg identified 45 universals on a sample of 30 languages which does not share common origins and geographical proximity. Greenberg analyzed many facets of language structure. First of all he focused on word order and set his investigation on the basis of the following criteria:

1) the existence of prepositions and postpositions (Pr/Po),
2) the relative order of subject, verb and object in declarative sentences with nominal subject and object
3) the position of qualifying adjectives in relation to nouns


First of all Greenberg based his investigation on the normal word order, in which normal means unmarked, of the main constituents of a sentence: subject (S), verb (V) and direct object (O). There are six possible orders: SVO, SOV, VSO, VOS, OSV, OVS. The most frequent are VSO, SVO and SOV (called I, II and III for the verb position). Romance languages (e.g. Italian), Germanic languages (e.g. English), Slavic languages, Finnish, Vietnamese, Swahili and many others have SVO order. Turkish, Hungarian, Japanese, Tamil etc have SOV order while Arabic, Welsh etc have VSO order. Let’s give some example:

Kataba muhammadun kitaban (Arabic)
V S O
La ragazza legge il libro (Italian)
S V O
The girl is reading the book (English)
S V O
Puella librum legit (Latin)
S O V


As we can notice, the subject always precedes the object. This is the fundamental assumption of the first universal:

Universal 1: In declarative sentences with nominal subject and object, the dominant order is always one in which the subject precedes the object.

The first universal implies a correlation between the basic word order in a sentence and the constituents of other grammatical constructs. Actually Greenberg found out that 14 languages out of 30 have postpositions and the genitive is followed by the governing noun while 15 have prepositions and the genitive after the governing noun. Only Norwegian can have the genitive before or after the noun. This is the notion expressed in:
Universal 2:
In languages with prepositions the genitive almost follows the governing noun (NG), while in languages with postpositions it almost always precedes (GN).

Greenberg employs the terms preposition (Pr) and postposition (Po) with the meaning of postmodification and premodification respectively. This gives us the chance to observe that among the three basic orders (VSO, SVO, SOV) we have:

- VO languages (VSO and SVO) in which the object follows the verb. They are postmodified (this means that the modifier comes after the V which is the head of the verbal phrase) and prepositional (this means that the head comes before the modifier)
- OV languages (SOV) in which the object precedes the verb, it stands on the left of the verb. These languages are premodified and postpositional.

Moreover Greenberg, looking at word order, noticed that there were several implications for the languages of the first, second and third type:

Universal 3: languages with dominant VSO(I) are always prepositional (NG – NA).

This implies that SOV(III) is the opposite of the I type and has statistically:
- postposition
- GN and AN order.

In many cases the second statement is not confirmed by empirical data because we can find G in the second position of the phrase (in other words NG) and when it occurs the adjective does likewise (NA):

Universal 4: with overwhelming greater than chance frequency, languages with normal SOV order are postpositional

Universal 5: if a language has dominant SOV order and the genitive follows the governing noun (NG), then the adjective likewise follows the noun (NA). This suggests that in many cases postpositions and the word order of other constructs may be disharmonic.


Frequently VSO languages put adverbs before the verb hence a sentence does not start with a verb. In addition VSO languages can present alternative word orders in which the most frequent is SVO while SOV languages put the verb after adverbs and other modifiers, and can present a OSV order. In both cases, SOV and OSV have the verb in the final position. Greenberg said that languages which had the verb in the last position were known as the rigid subtype. These considerations represent the universals 6 and 7:

Universal 6: all languages with dominant VSO order have SVO as alternative or as the only alternative basic order.
This suggests that VSO order is in some sense a variant of SVO order

Universal 7: if in a language with dominant SOV order there is no alternative basic order or only OSV as alternative then all adverbial modifiers of the verb likewise precede the verb
This is pretty clear: if V is final, it’s final.

Greenberg investigated syntax of his sample and in particular focused on interrogative sentences which can be:
1)yes/no questions
2)questions involving specific words

We know that a common hint to distinguish interrogative questions from statements is related to intonation. In English for example a yes/no question is marked with a rise in pitch in the last stressed syllable of the sentence while the corresponding statement is marked with a fall in pitch:

Universal 8: when a yes/no question is differentiated from the corresponding assertion by an intonational pattern the distinctive features of each of these patterns are reckoned from the end of the sentences rather than from the beginning

We have to say that a yes/no question can be rendered not only thanks to intonation but also to question particles or affixes which usually are related to the verb or the emphasized word. In subtype III it is impossible to distinguish the position after the verb and the final position in the sentence. Greenberg found out that 5 VSO languages use initial particles, while 2 SVO and 5 SOV final particles. This implies that when the interrogative particles are initial they occur in prepositional languages, while when they are final they occur in postpositional languages:

Universal 9: With well more than chance frequency, when question particles or affixes are specified in position by reference to the sentence as a whole, if initial, such elements are found in prepositional languages, and, if final, in postpositional.

Greenberg underlines that the question marker follows the emphasized word of the question, this phenomenon occurs in languages of the type II and III but not in I:

Universal 10: Question particles or affixes, when specified in position by reference to a particular word in the sentence, almost always follow that word. Such particles do not occur in languages with dominant order VSO.

As said above, there is another kind of questions: questions involving specific words. In this kind of sentences many languages employ a different word order than that of the corresponding statement. Usually in many languages the question word and the question phrase come first and the verb precedes the subject. In the same languages there is also an inversion for yes/no questions. Moreover VSO languages put the question word first, SOV ones have the same word order for statements and questions while SVO ones statistically act as VSO ones but in two cases as SOV ones:

Universal 11: Inversion of statement order so that verb precedes subject occurs only in languages where the question word or phrase is normally initial. This same inversion occurs in yes-no questions only if it also occurs in interrogative word questions.
This suggests a structural relationship between the position to which the verb is inverted and the initial position that wh elements occupy.

Universal 12: If a language has dominant order VSO in declarative sentences, it always puts interrogative words or phrases first in interrogative word questions; if it has dominant order SOV in declarative sentences, there is never such an invariant rule.
An initial position for wh elements (of course combined with prosodic features) makes functional sense, since it means that the hearer knows at the beginning that he/she is listening to a question. Similarly an early position of the verb is useful to produce a question.

To sum up, in prepositional languages tendentially we have the following orders: SO, NG, NA, initial question particles and initial interrogative words or phrase.

Greenberg focused on the relative order of subordinate and main verbal forms. We know that a subordinate verb qualifies the main verb and we could deduce that the subordinate always precedes the main verb in all languages of the rigid subtype III. The empirical investigation endorsed this deduction and actually in all languages of the rigid subtype III the subordinate verb comes before the main one. Moreover Greenberg found out that in all 30 languages the protasis of conditional sentences precedes the apodosis but in some cases can be an inversion while in the subrigid type III the protasis always precedes the apodosis. On the contrary the expressions of purpose and volition statistically follow the main verb except for the subrigid type III in which precede the main verb.


These considerations are explained in the following universals:

Universal 13: if the nominal object always precedes the verb then verbs forms subordinate to the main verb also precede it
This suggests that it happens in SOV which is usually postpositional, hence the verb is the head and its modifiers (object and subordinate forms) come before the head.

Universal 14: in conditional clauses, the conditional clause (protasis) precedes the conclusion (apodosis) as the normal order in all languages
In this case the sequence protasis – apodosis can be considered the unmarked form or if you prefer we can say: ‘if x then y’

Universal 15: in expression of volition and purpose, a subordinate verb always follows the main verb as the normal order except in those languages in which the nominal object always precedes the verb (subrigid type III)
In strongly head-final languages every element precedes the main verb.

Greenberg found out that in prepositional languages the inflected auxiliary precedes the verb while in postpositional ones follows it. Of course, Greenberg excludes languages in which the auxiliary is not inflectional (e.g. Japanese); 19 languages out of 30 have an inflectional auxiliary:

Universal 16: in languages with dominant order VSO an inflected auxiliary always precedes the main verb. In languages with dominant order SOV an inflected auxiliary always follows the main verb
In VSO languages which are prepositional the auxiliary is the head hence it comes before the full verbal form.

This universal implies a comparable word order in the relationship between qualifying adjectives and nouns. Greenberg found out that in prepositional languages statistically adjectives follow nouns (NA) while in postpositional languages precede them:

Universal 17: with overwhelmingly more than chance frequency languages with dominant order VSO have the adjective after the noun
If a language has the sentential head verb at the beginning, it has the head noun at before the adjective.


To sum up, in prepositional languages tendentially we have the following orders: SO, NG, initial question particles, initial interrogative words or phrase, Prot.Apod., AuxV and NA.

This last universal has two additional implications:
1) in postpositional languages when a qualifying adjective precedes the noun then demonstratives and numerals do the same
2) usually in prepositional usually the order is NA but demonstratives and numerals statistically tend to act in a different way, in particular demonstratives tend to precede the noun while numerals tend to follow it:

Universal 18: when a descriptive adjective precedes the noun the demonstrative and the numeral overwhelmingly more than chance do likewise
This is purely implicational: if AN then DemN then NumN.

Universal 19: when the general rule is that the descriptive adjective follows there may be a minority of adjectives which usually precede but when the general rule is that descriptive adjectives precede there are no exceptions
If AN no exceptions, if NA several exceptions.

Moreover Greenberg underlines that when we have at the same time qualifying, demonstrative and numeral adjectives before the noun they are placed with the following order:
1 2 3 4
- demonstrative, numeral, adjective + noun,


while when the adjectives follow the noun we have the opposite fixed order:

4 3 2 1
- noun + adjective, numeral, demonstrative. This is stated by:

Universal 20: When any or all of the items (demonstrative, numeral, and descriptive adjective) precede the noun, they are always found in that order. If they follow, the order is either the same or its exact opposite

Adverbs acts toward adjectives like adjectives toward nouns and auxiliary toward the main verb. Of course, we must keep in mind that the adverb is the modifier and the adjective the head. Greenberg noticed that in all of those languages in which there is NA, there is also AAdv and they are of the first and second type:

Universal 21: If some or all adverbs follow the adjective they modify, then the language is one in which the qualifying adjective follows the noun and the verb precedes its nominal object as the dominant order
AAdv is related to NA and VO (VSO and SVO).

We know that comparative adjectives can be expressed for example in English with:
- an inflected comparative form of the adjective (in English, A is smaller than B)
- a separate word which modifies the adjective (in English, A is more stupid than B)

This second option is not so frequent in languages. Usually the second term of comparison (e.g. than + the item we are comparing in English) is expressed in many languages with affixes or specific words. For example in English ‘larger than B’ larg is the adjective, er the marker of comparison and than the standard of comparison. Then we can find two possible sequences:

1) adj (A)-marker (M)-standard (S) which is typical of prepositional languages
2) standard-marker-adj, typical of postpositional languages

These two options represent the:

Universal 22: If in comparisons of superiority the only order, or one of the alternative orders, is standard-marker-adjective, then the language is postpositional. With overwhelmingly more than chance frequency if the only order is adjective-marker-standard, the language is prepositional

To sum up, in prepositional languages tendentially we have the following orders: SO, NG, initial question particles, initial interrogative words or phrase, Prot.Apod., AuxV, NA, AAdv and AMS.


Another implication related to word order typology regards apposition, in particular a common noun with a proper noun. The investigation led Greenberg to conclude that:

Universal 23: If in apposition the proper noun usually precedes the common noun, then the language is one in which the governing noun precedes its dependent genitive. With much better than chance frequency, if the common noun usually precedes the proper noun, the dependent genitive precedes its governing noun

In other words NG implies Proper/Common while GN implies Common/Proper, the proper noun is the head and it is found in prepositional languages whereas it comes before the common noun.


We know that a relative clause modifies a noun. Greenberg found out that relative clauses (RC) precede the noun in postpositional languages while they follow it in prepositional ones. Anyway Chinese is an exception because it is prepositional but its relative clauses always precede the noun:

Universal 24: If the relative expression precedes the noun either as the only construction or as an alternate construction, either the language is postpositional, or the adjective precedes the noun or both.
We can summarize as follows: if RCN then postposition then AN. In all these cases the head follows a dependent.

To sum up, in prepositional languages tendentially we have the following orders: SO, NG, initial question particles, Prot.Apod., AuxV, NA, AMS, Proprer/Common and NRC.


The last feature investigated by Greenberg was the pronominal order. In many languages such as Italian, French, Greek Swahili the pronominal object always precedes the verb (PronV) while the nominal object follows (Je la prends/Je prends une aspirine) but if we have an imperative form then the pronoun and the object come after the verb (Prendsla/ prends une aspirine). In Berber we have PronV whereas the verb is in negative or future form. Statistically it’s frequent that the pronominal object precedes the verb while the nominal object can come before or after the verb. In VSO languages (statistically prepositional) the pronominal object immediately follows the verb and precedes the subject (the sequence is V Pron. S) and this is in perfect harmony with prepositional languages in which the verb is the head of the verbal phrase (VPron):

Universal 25: If the pronominal object follows the verb, so does the nominal object
If VPron then VO.


Greenberg analyzed his sample from a threefold point of view: word order, syntax and morphology. The universals 26 – 45 concern morphology. He gave a quick explanation of morphemes from a functional point of view, hence he listed them as follows:

- root
- derivational
- inflectional
- derivational and inflectional ones are grouped as affixes
- affixes can be prefixes, suffixes and infixes
- suffixes are the most frequent in languages followed by prefixes and infixes
- many languages are not completely suffixes or prefixes oriented but can present both phenomena, hence:

Universal 26: if a language has discontinuous affixes, it always has either prefixing or suffixing or both
Of course a full explanation of morphemes is not present in Greenberg’s article but we know that morphemes can be divided in two major classes: lexical morphemes and grammatical morphemes which are subdivided in derivational and inflectional morphemes.

Greenberg got the empirical evidence that:

Universal 27: if a language is exclusively suffixing, it is postpositional; if it is exclusively prefixing, it is prepositional

This suggests there is a relation between morphology and syntax. If we look at Greenberg’s 30 languages sample we find that 12 languages are purely suffixing and are postpositional while 17 languages have prefixes as well as suffixes: 15 of them are prepositional, 2 postpositional. Only one language is exclusively prefixing and is prepositional. Hence the first part of this universal is well generalized but the second one, in my opinion, is based only on a single example and the generalisation is more to difficult to do.

To sum up, in prepositional languages tendentially we have the following orders: SO, NG, initial question particles, Prot.Apod., AuxV, NA, AMS, NRC, VPron and prefixes or prefixes and suffixes.



Whereas in languages grammatical morphemes of two subtypes are present and follow or precede the root, we have the following sequence: root + derivational m. + inflectional m. or inflectional m. + derivational m. + root. For example in Italian we have: cambiamento which can be divided in:
Cambi –ament –o
root deriv. infl.

This is the assumption stated by:

Universal 28: If both the derivation and inflection follow the root, or they both precede the root, the derivation is always between the root and the inflection

Moreover Greenberg said that there are not languages which are purely isolating, in other words languages where the word structure is very simple(we have only the root) and to one word corresponds one morpheme, usually we have languages with compounding, affixes or both:

Universal 29: If a language has inflection, it always has derivation

This suggests that Greenberg described the relation between inflection and derivation in distributional terms. Of course this universal means that if a language presents phenomena of inflection it presents phenomena of derivation in the same or different words (gatt-o gatt-in-o).
Universal 29 implies that languages without inflection do not present verb inflectional categories:

Universal 30: if the verb has categories of person-number or if it has categories of gender, it always has tense-mode categories

This is pretty clear, tenses and modes in verbs are expressed by inflectional morphemes.

The following universals are based on the agreement in gender and number between the subject and the verb and the nominal object and the verb:

Universal 31: If either the subject or object noun agrees with the verb in gender, then the adjective always agrees with the noun in gender

Universal 32: Whenever the verb agrees with a nominal subject or nominal object in gender, it also agrees in number

In many cases the agreement between the verb and the noun is suspended in particular when we have VS or VO and the verb is singular:

Universal 33: When number agreement between the noun and verb is suspended and the rule is based on order, the case is always one in which the verb precedes and the verb is in the singular

We know that gender (M/F), number (Sing/Plur) and case (Nom/Gen/Dat etc) are the typical inflectional categories. These three categories have implications among them:

Universal 34: No language has a trial number unless it has a dual. No language has a dual unless it has a plural, (e.g. ancient Greek has singular, plural and dual)

Universal 35: There is no language in which the plural does not have some nonzero allomorphs, whereas there are languages in which the singular is expressed only by zero. The dual and the trial are almost never expressed only by zero.
In this last case the singular represent the unmarked category while the others the marked ones.

The interralation between gender and number are explained in the following universals:

Universal 36: If a language has the category of gender, it always has the category of number.
Another implication: if gender then number.

Universal 37: A language never has more gender categories in nonsingular numbers than in the singular.
In other words the plural can have the same number of gender categories of the singular or less than the singular and Greenberg in order to endorse this statement says that in Hausa we have M/F gender distinction in the singular but not in the plural (the plural has less gender categories than the singular). I’m not so sure about this. I remember in some Austronesian languages the plural makes distinction between animated and non animated forms while such distinction is not present in the singular.

In relation to cases, Greenberg found out that:

Universal 38. Where there is a case system, the only case which ever has only zero allomorphs is the one which includes among its meanings that of the subject of the intransitive verb

In many languages, called ergative, with cases system the mark of the subject varies according to the transitive or intransitive nature of the verb. While in Greek, Latin, Russian we have the classical nominative/accusative, here we have absolutive/ergative, in which the absolutive is the case of the direct object of transitive verbs but also of the subject of intransitive verbs and the ergative is the case of the subject in transitive sentences.

Moreover if we consider:
Vas vekerula (the boy runs)

Vas v eker ula
The boy masculine to run simple present
Abs Sing. Abs.

Vas is the absolutive case and it is formed by the ‘naked form of the word’ i.e. there is not a case morpheme, this is understandable if we consider the absolutive the unmarked case.


Universal 39: Where morphemes of both number and case are present and both follow or both precede the noun base, the expression of number almost always comes between the noun base and the expression of case

Of course a simple empirical evidence of this last statement is not easy to find especially when we must deal with unknown languages, anyway I will try to give an example from Armenian:

Enker-ner-ic : by comrades
Enker is the root
Ner is the plural
Ic case ablative

Universal 40: When the adjective follows the noun, the adjective expresses all the inflectional categories of the noun. In such cases the noun may lack overt expression of one or all of these categories.

In other words when we have NA the adjective carries gender, number and possibly case while the noun does not. This is very common in Basque.

As a general statement, Greenberg maintains that cases are frequent in postpositional languages (Basque, Turkish, Japanese) and in particular in SOV ones, hence:


Universal 41. If in a language the verb follows both the nominal subject and nominal object as the dominant order, the language almost always has a case system
If postpositional then case system.

The last four universals deals with pronominal categories:

Universal 42. All languages have pronominal categories involving at least three persons and two numbers.
They are masculine, feminine, neuter, singular, plural


Universal 43. If a language has gender categories in the noun, it has gender categories in the pronoun
The pronoun acts in substitution of a noun.

Universal 44. If a language has gender distinctions in the first person, it always has gender distinctions in the second or third person, or in both.
This suggests that gender oppositions in personal pronouns are characteristic of the third rather than the first or second person.



Universal 45. If there are any gender distinctions in the plural of the pronoun, there are some gender distinctions in the singular also.

If M/F plural then M/F singular, gender is seen as being typical of singular rather than non-singular personal pronouns.

Conclusions
From the scanning of his sample Greenberg noticed that while the nominal object could follow the verb the pronominal object instead could follow or precede it, hence he underlined that the order VO is dominant over OV because the latter can be found only when the pronominal object precedes the verb while VO does not have such limitation. Moreover Greenberg discovered that the order OV is in harmony with PronV but not with VPron. In the same way VO is in harmony with VPron but not with PronV. As said above, the harmony can be explained in terms of premodification and postmodification. These considerations led Greenberg to formulate the following statement: a dominant order may always occur, but its opposite, the recessive, occurs only when a harmonic construction is likewise present. In other words: if OV then PronV, if VO then PronV. The notions of harmony and disharmony are used by Greenberg in terms of generalization. Moreover he analyzes universal 3 which states that SOV languages are always prepositional and in the three main word orders we notice that in two cases the subject precedes the verb (SV) while in one follows. Hence Greenberg maintains that prepositions are dominant over postpositions (the recessive alternative occurrence is zero) and prepositions are in harmony with VS and in disharmony with SV while postpositions are in harmony with the opposite order, i.e. SV
and not with VS. Universal 4 states that SOV is associated with postpositions. This means that OV is in harmony with postpositions while VO (I and II) with prepositions. In addition we can notice:
- in VSO (I type): VS and SO are harmonic with prepositions (in VS, V is the head, In SO, S is the head) and prepositions are dominant in other constructs
- in SVO (II type): SV is harmonic with postpositions and VO with prepositions and prepositions are dominant in other constructs
- in SOV (III type): we found SV and OV which are in harmony with postpositions but we have to say that in this case many SOV languages are postpositionals with several exceptions.
Going this way we see that NG is in harmony with prepositions while GN with postpositions. This implies that NA and AN (the adjective ‘defines’ like the genitive) are associated with prepositions and postpositions. In other words we find that in prepositions we have the following orders: NG, VS, VO, NA while in postpositions we have the opposite swapped order: GN, SV, OV, AN. In many cases (universal 5) we can have a disharmony between the Prep/Post and the word order of other constructs. In other words we can have languages of the type III with prepositions but also languages of the type II with prepositions and with NG or GN and NA and one type I with GN and AN. In universals 20 and 29 Greenberg found the so called proximity hierarchies in which the main element is the root morpheme and there is an implicational hierarchy in the case of inflection. Moreover number categories are closer to the root (generating an inflection) than the case. The notion of proximity is related to the neutralization one. Hence there are categories which depends on (are the implication of, are neutralized by) other. Number is neutralized by gender (universal 36), gender in non singular are neutralized by singular ones (universal 37), trial depends on dual and so on (universal 34). Moreover Greenberg noticed that universal 7,8,40 have in common the idea of marking the end of a unit or a sentence.