Newmeyer Handout #2

1

4. TYPOLOGY, UNIVERSAL GRAMMAR, AND PARSING

Prep 6 10 0 Postp 0 3 11

Correlations between word orderand adposition order (Greenberg 1963)

Table 1

(1) UG characterizes the notion ‘possible human language’, not the notion‘probable human language’.

There has also been very productive study of generalizations that aremore directly observable: generalizations about the word orders weactually see, for example. The work of Joseph Greenberg has beenparticularly instructive and influential in this regard. These universals areprobably descriptive generalizations that should be derived from principles of UG.(Chomsky 1998: 33; emphasis added)

[W]hat we ‘know innately’ are the principles of the various subsystems ofS0 [= the initial state of the language faculty — FJN] and the manner oftheir interaction, and the parameters associated with these principles.What we learn are the values of these parameters and the elements of theperiphery (along with the lexicon, to which similar considerations apply).The language that we then know is a system of principles with parametersfixed, along with a periphery of marked exceptions. (Chomsky 1986: 150-151)

it would make sense if children, too, instinctively work their way downthe hierarchy, taking advantage of its logical structure to avoid agonizingover needless decisions (Baker 2002: 192)

(2) W1 - W2 - W3 - ... - Wn —> Wn - ... - W3 - W2 - W1

(3) Guiding assumption in mainstream generative work: there is no disparitybetween the notions ‘typologically significant generalization’ and ‘linguistically(i.e. grammatically) significant generalization’

Chinese is consistently head final except in rule expanding X’ to X0 (if the head isverbal it precedes the complement)(4) (Huang 1994):

a. XP —> YP X'b. X' —> YP X'c. X' —> c'. X0 YP iff X = [+v]

c''. YP X0 otherwise

Newmeyer Handout #2

2

(5) CP (Kayne 1994)

[spec, CP] C'

IP C IPt

(6) *NDem & NumN predicted by KayneBut Berber, Welsh, Hebrew, Zapotec , etc. have this correlationThese languages have and extra movement of demonstratives (Cinque

1996)

(7) Possible orders of V, direct object, complement PP1, and adjunct PP2:a. PP2 PP1 NP V e. PP1 NP V PP2b. PP2 PP1 V NP f. PP1 V NP PP2c. PP2 NP V PP1 g. NP V PP1 PP2d. PP2 V NP PP1 h. V NP PP1 PP2

HEADEDNESS THETA CASE LANGUAGE

a. final left left Japaneseb. final left right Chinese (future)c. final right left *d. final right right Chinese (present)e. initial left left Kpelle (past)f. initial left right *g. initial right left Kpelle (present)h. initial right right English

Combinations of the headedness, direction of theta-role assignment, anddirection of case assignment parameters (Travis 1989)

Table 2

For specific proposals concerning marked values to entail testable claims,these claims will have to hold in an ‘external’ domain, a domain otherthan that of the distribution of morphemes or grammatical well-formedness. Claims to explanatory adequacy will have to be grounded insuch domains. Natural candidates for such a domain wherein markednessproposals make empirically testable claims are language change and acquisition.(Lightfoot 1979: 76-77; emphasis added)

[In]n determining which notions are encoded in a language’s morphology,the child is faced with a formidable search problem ... [B]y imposing aweighting on the child’s hypotheses, one could account for the largedisparities in the prevalence of various grammatical encodings in theworld’s languages, and in the speed of acquisition of various encodings bychildren. (Pinker 1984: 168-171)

Newmeyer Handout #2

3

One intriguing possibility is that the relative accessibility for children ofalternative schemes for partitioning meaning in a given conceptualdomain is correlated with the frequency with which these schemes areinstantiated in the languages of the world. ... It is plausible that relativefrequency is correlated with ‘ease’ or ‘naturalness’ for the human mind.(Bowerman 1985: 1306).

(8) a. Cross-linguistically frequent properties of language are reflected bycorrespondingly simple (unmarked) properties of grammars.

b. Cross-linguistically frequent properties of language are acquired earlyby the child.

c. Cross-linguistically frequent properties of language are diachronicallystable.

If (8a-c) were right, then typology would be relevant to gram theory in twocomplementary ways:

1.One could appeal to grammatical theory to explain typologicaldistribution

2.Typological distribution would be a reliable heuristic for grammaticalanalysis

But they are not right!

Emonds 1980: Irish and other VSO languages are underlyingly SVO: they haverule that fronts the verb. Their grammars are ‘more complicated, therefore rarer’(p. 44).

Baker 2001: the more ‘choices’ a language learner needs to make on theParameter Hierarchy, the rarer the language type is claimed to be.

(9) a. Who did you talk to?b. Mary was spoken to.

Stranding grammars are simpler than non-stranding grammars:(10) a. NON-STRANDING LANGUAGES: The lexical categories N, V, and Aare proper governors. The lexical category P is not a proper governor.

b. STRANDING LANGUAGES: All four lexical categories are propergovernors.A popular reanalysis approach to stranding (Hornstein & Weinberg 1981)(11) a. You talked PP[to who] > You V[talked to] who > Whoi did you V[talk to] ei? b. e was spoken PP[to Mary] > e was V[spoken to] Mary > Maryi was V[spokento] ei

But reanalysis demands implausible lexical items:(12) a. Which shoes did you [walk across Europe in] ___? (Jones 1987)

b. Which of the two knives did you [pay twice for] ___? (Inada 1981)

Newmeyer Handout #2

4

Gapping doesn’t treat the V-P complex as a verb:(13) a. *John looked at Mary and Bill ___ Sue.

b. John looked at Mary and Bill ___ at Sue.

Nor does Heavy NP-shift:(14) a. John looked at [the woman he loved] very often.

b. John looked very often [at the woman he loved]c. *John looked at very often [the woman he loved].

Extraposition out of a lexical item would be necessary:(15) What did you [talk to that guy ___ about] ___who was here yesterday?

Mutually incompatible analyses, where wh-movement and passive both apply:(16) a. Which problems has Harry been [[talked to] e about] e?

b. Who would you like to be [[sung to] e by] e?

If P is a proper governor, then V and P don’t have to be adjacent:(17) a. Who did you give all those books about golf to?

b. Which burner did you leave the pot on?

Most interesting prediction of this analysis: stranding is possible withextraction of NP from PP adjuncts to VP:

(18) VP

V' PP

V (NP) P NP [+wh]

This is correct:(19) a. Which shoes did you walk across Europe in?

b. Which ball park did Ruth hit the most home runs in?c. Which knife shall we use to cut the turkey with?d. Which red-headed man is Mary standing beside?

See Bayer (1999) on problems with Kayne’s treatment of COMP-final language’slack of Wh-movement

V-final SVO V-initialWh-in situ 71 42 16Final Q particles 73 30 13

Proportion of languages with Wh-in situ and final question particles,by word order type (Dryer 1991)

Table 3

Newmeyer Handout #2

5

(20) Polysynthesis

no yes

Head Directionality—Optional Polysynthesis Adjective Neutralizeverb noun

first/no * * Mohawk Warlpiri

first/yes last/yes last/no

Subject Side * * Ergative CaseChichewa Slave Accusative ErgativeSelayarese Quechua

Topic Prominent *beginning Greenlandic

end yes no Dyirbal

Verb Attraction * * * Tzotzil Japanese Turkish Malagasy Choctaw Malayalam

yes no

Subject Placement Serial Verbs

low high no yes

* Null Subject * *Welsh English EdoZapotec Indonesian Khmer

no yes* *

French SpanishRomanian

The Parameter Hierarchy (Baker 2001: 183)

V-initial V-medial V-final42 30 64

Percent of languages of each type with explicit dependent (case) marking(Siewierska and Bakker 1996)

Table 4

(21) *the man met you ___ is my friend(22) *[NPNP tense VP]

Newmeyer Handout #2

6

[I]t is difficult to believe that everyone who knows [this filter] has beenexplicitly corrected (or has observed corrections) for violation. Dismissingthis possibility, we must assign to the language faculty some principle thatleads to postulating [the presence of this filter] as the ‘unmarked case’unless there is specific evidence to the contrary. (Chomsky and Lasnik 1977:437)

(23) [NPthe man own the land] come over

(24) There is little or no correlation between age of acquisition and typologicalfrequency:

a. American children acquire P-stranding before pied-piping (KarinStromswold, personal communication)

b. French-speaking children have the verb-raising parameter from theearliest multi-word utterances (Pierce 1992)

c. there is no period during which German-speaking children fail to set theV2 parameter (Poeppel and Wexler 1993)

d. the null subject parameter is set very early, regardless of whether thelanguage is null subject or not (Valian 1991)

e. children acquiring English, German, and French evidence strongknowledge of locality in wh-extraction domains at early ages (Roeper and DeVilliers 1994)

(25) The Strong Continuity Hypothesis: The parameters of UG are set earlyand are not subject to change in the course of development (see especially Lust,Suñer and Whitman 1994)

This hypothesis is probably too strong, however (Clahsen and Penke 1992;Radford 1996; Clahsen, Eisenbeiss, and Penke 1996)

(26) Early Immediate Constituents (EIC) (Hawkins 1994)The human parser prefers linear orders that maximize the IC-to-non-ICratios of constituent recognition domains (CRD).

(27) a. ?I consider everybody who agrees with me and my disciples about thenature of the cosmos (to be) smart.

b. I consider (to be) smart everybody who agrees with me and mydisciples about the nature of the cosmos.(28) a. ?I met the twenty three people who I had taken Astronomy 201 withlast semester in the park.

b. I met in the park the twenty three people who I had taken Astronomy201 with last semester.

Newmeyer Handout #2

7

(29) a. S b. S

NP VP NP VP

I V NP PP I V PP NP

met D N’ P NP met P NP D N’

the 23 ... 201 in the park in the park the 23 ... 201

(30) a. That John will leave is likely.b. It is likely that John will leave.

(31) a. Mary-ga [kinoo John-ga kekkonsi-ta to] it-ta Mary yesterday John married that said

‘Mary said that John got married yesterday’b. [kinoo John-ga kekkonsi-ta to] Mary-ga it-

(32) a. S[S’[ that S[ John will leave]] VP[[is likely]]

b. S[ NP [it] VP[ is likely S’[that S[John will leave]]]]

(33) a. S1[Mary-ga VP[S’[S2 [kinoo John-ga kekkonsi-ta] to] it-ta]]

(34) a. VP b. VP

V NP PP PP NP V

P NP NP P

c. VP d. VP

V NP PP PP NP V

NP P P NP

(35) a. [N Adj [C S]] g. [N [C S] Adj]b. [Adj N [C S]] h. [Adj [C S] N]c. [[S C] N Adj] i. [N [S C] Adj]d. [[S C] Adj N] j. [Adj [S C] N]e. [N Adj [S C]] k. [[C S] N Adj]f. [Adj N [S C]] l. [[C S] Adj N]]

Newmeyer Handout #2

8

(36) a.’ NP e.’ NP

N Adj S’ N Adj S’

C S S C

g.’ NP k.’ NP

N S’ Adj S’ N Adj

C S C S

The Prepositional Noun-Modifier Hierarchy (PrNMH) of Hawkins 1983:(37) PrNMH

If a language is prepositional, then if RelN then GenN, if GenN thenAdjN, and if AdjN then DemN.

(38) a. PP[P NP[___N…] (Arabic, Thai)b. PP[P NP[___N…]; PP[P NP[Dem N…] (Masai, Spanish)c. PP[P NP[___N…]; PP[P NP[Dem N…]; PP[P NP[Adj N…] (Greek, Maya)d. PP[P NP[___N…]; PP[P NP[Dem N…]; PP[P NP[Adj N…];

PP[P NP[PossP N…] (Maung)e. PP[P NP[___N…]; PP[P NP[Dem N…]; PP[P NP[Adj N…];

PP[P NP[PossP N…]; PP[P NP[Rel N…] (Amharic)

The Greenbergian correlations hold better at surface level than at deep level.

(39) German and Dutch:a. main clause declaratives: SVOb. subordinate clauses: SOV

But German and Dutch are not typologically SOV

(40) SURFACE ORDER AND TYPOLOGICAL GENERALIZATIONSa. Dutch and German are more ‘consistent’ at the surface than

underlyingly.b. Despite Bach (1970a), Amharic is typologically more like an OV than a

VO language.c. Languages with discourse-governed word order (where it is hard to

make a case for one or another deep order of elements) manifest the correlationsassociated with that surface order having the highest text frequency (see Table 5).

d. Some languages have been argued by generativists to have nounderlying order of subject, object, and verb (Kiss 1987 for Hungarian, Hale 1992

Newmeyer Handout #2

9

for Warlpiri, etc.) Clearly, in such languages the Greenbergian correlationscannot be captured at D-structure.

Ute 72% GenN, PoTlingit 67% GenN, Po, RelNHuallaga Quechua 69% GenN, Po, RelNTrumai 65% GenN, PoKoryak 66% GenN, PoTacana 86% GenN, Po, Clause-final subordinatorTakelma 85% GenN, PoHupa 53% GenN, PoCherokee (more frequent) GenN, PoKorana 89% GenN, Po, Clause-final Q, etc.

Frequency of OV order and OV characteristics among languages with purelydiscourse-governed orders (Dryer 1989)

Table 5

(41) Transformational theories in which ordering of elements is a low-levelphenomenon: Sanders 1970, Staal 1967; Peterson 1971, Hudson 1972; Chomsky1995.

Language design appears to be problematic from a parsing-theoreticperspective, though elegant regarded in isolation from considerations ofuse. (Chomsky 1991: 448)

Property V-final SVO V-initialPostpositional 96 14 09Relative-Noun 43 01 00Standard of comparison-Adjective 82 02 00Predicate-Copula 85 26 39Subordinate clause-Subordinator 70 06 06Noun-Plural word 100 24 13Adpositional phrase-Verb 90 01 00Manner Adverb-Verb 91 25 17Verb-Tense/aspect aux verb 94 21 13Verb-Negative auxiliary 88 13 00Genitive-Noun 89 59 28Sentence-Question particle 73 30 13Wh-in situ 71 42 16

Table 6Percent of V-final, SVO, and V-initial languages manifesting particular properties

(Dryer 1991)

Newmeyer Handout #2

10

(42) Occurrence of distinctive reflexives:

3rd person 1st /2nd PersonEnglish yes yesOld English no noFrench yes no

* no yes

Faltz and Comrie have argued that the data in (42) result from it being more‘useful’ to have 3rd person reflexives than 1st or 2nd, since that helps to reduceambiguity

DIFFERENTIAL OBJECT MARKING (DOM): The higher in the hierarchies ofanimacy and/or definiteness (see 2 and 3) a direct object is, the more likely it is tobe case marked in a particular language:

(43) Animacy Hierarchy: Human > Animate > Inanimate(44) Definiteness Hierarchy: Personal Pronoun > Proper Noun > Definite NP >

Indefinite Specific NP > Non-specific NP

NP +def -def +pron -pron +anim -animSubj 3151 3098 53 2984 167 2948 203Obj 3151 1830 1321 1512 1639 317 2834

Table 7Frequencies in the SAMTAL corpus of spoken Swedish (Jäger 2003)

An intuition which recurs in the literature on DOM is that it is those directobjects which are most in need of being distinguished from subjects that getovertly case marked. This intuition is sometimes expressed as the idea that thefunction of DOM is to distinguish subject from object. … In a weaker form, theintuition can be understood in the following terms: the high prominence whichmotivates DOM for objects is exactly the prominence which is unmarked forsubjects. Thus it is those direct objects which most resemble typical subjects thatget overtly case marked. (Aissen 2003)

Reflexive pronoun Number ofoccurrences in corpus

myself 169yourself 94himself 511herself 203itself 272TOTAL 3RD PERS. SG. 986

Table 8: Reflexive Pronoun Occurrence in a million-word collection of Englishtexts (Johansson and Hofland 1989)

Newmeyer Handout #2

11

(45) Johansson and Hofland 1989:a. I 7620 myself 169 = 2.2%b. he 9068 himself 511 = 5.6%

Reflexive pronoun Number in corpusmyself 151yourself 189himself 129herself 43itself 296TOTAL 3RD PERS. SG. 468

Table 9Reflexive Pronoun Occurrence in the 1,848,364 word Michigan Corpus of AcademicSpoken English

The drive to reduce formal ambiguity cannot be an important force in languagedesign and use because the transmission rate of human speech is so slow

Sentence (46) was calculated to have 455 parses:

(46) List the sales of products produced in 1973 with the products produced in1972

[I]nference is cheap, articulation expensive, and thus the designrequirements are for a system that maximizes inference. (Levinson 2000:29)

Grammars provide the most economical coding mechanism … for thosespeech functions which speakers most often need to perform. Moresuccinctly: Grammars code best what speakers do most. (Du Bois 1985: 362-363)

(47) a. Only a minority of languages make an inclusive-exclusive pronoundistinction (Nichols 1992).

Washo:

Sg. Dual Pl.1st exclusive lé> lés&i (= I & one other) léw (= I & others)1st inclusive lés&is&i (= I & you [sg.]) léwhu (= I & you [pl.])

b. Second person singular and plural pronouns have fallen together inmany European languages.

c. The majority of the world’s languages are null subject (Gilligan 1987).

Newmeyer Handout #2

12

(48) a. Very few deictic systems encode the height of an object in relationto the speech participants

b. No deictic system encodes whether an object is to the left or theright of the speaker/hearer

A striking fact discovered by Len Talmy: No language has grammatical meansfor encoding an event structure that is independent of that encoded by the lexicalcategories in the sentence. That is, no language has sentences like (49) withmeanings like (50a-b):

(49) The chair broke-ka(50) a. The chair broke and I’m currently bored.

b. The chair broke and it was raining yesterday.

Ambiguity-reduction is not an important factor in language change.

Form-function reanalysis [one of Croft’s principal mechanisms of change— FJN] is syntagmatic: it arises from the (re)mapping of form-functionrelations of combinations of syntactic units and semantic components. Theprocess may nevertheless have an apparently paradigmatic result, for example, achange of meaning of a syntactic unit … (Croft 2000: 120)

Appeals to ambiguity-reduction are psychologically-implausible

Word order frequencies:(51) SVO/SOV > VSO >VOS/OSV > OSV

If speakers were really driven to reduce ambiguity, they would put the objectbefore the subject

It is easier to predict the properties of the direct object of a transitive sentencegiven the subject than to predict properties of the subject, given the direct object

So an ambiguity-reduction theory predicts objects to precede subjects

An ambiguity-reduction theory also makes the wrong predictions with regard togrammatical relations and case-marking

Subjects are the least likely grammatical relation to be overtly case-marked — themore ‘oblique’ the relation, the more likely case marking is

But case marking not only reveals properties of the element case-marked, but italso reveals properties of other sentential elements, in particular the predicate

So, a benefactive marker, say, on an NP restricts severely the class of verbspossible in the sentence, namely, to those that can take benefactive arguments.

An ambiguity-reduction theory predict incorrectly that benefactive NPs shouldin general occur earlier than (nonpredictive because non-case-marked) subjects

Newmeyer Handout #2

13

REFERENCES

Bayer, Josef. 1999. 'Final Complementizers in Hybrid Languages', Journal of Linguistics 35, 233-272.Bowerman, Melissa. 1985. 'What Shapes Children's Grammars?' in D. I. Slobin (ed.), The

Crosslinguistic Study of Language Acquisition, Erlbaum, Hillsdale, NJ, pp. 1257-1319.Chomsky, Noam. 1991. 'Some Notes on Economy of Derivation and Representation', in R. Freidin

(ed.), Principles and Parameters in Comparative Grammar, MIT Press, Cambridge, MA, pp.417-454.

Chomsky, Noam. 1995. The Minimalist Program, MIT Press, Cambridge, MA.Chomsky, Noam. 1998. 'Noam Chomsky's Minimalist Program and the Philosophy of Mind. An

Interview [with] Camilo J. Cela-Conde and Gisèle Marty', Syntax 1, 19-36.Chomsky, Noam and Howard Lasnik. 1977. 'Filters and Control', Linguistic Inquiry 8, 425-504.Cinque, Guglielmo. 1996. 'The 'Antisymmetric' Program: Theoretical and Typological

Implications', Journal of Linguistics 32, 447-465.Dryer, Matthew S. 1989. 'Discourse-Governed Word Order and Word Order Typology', Belgian

Journal of Linguistics 4, 69-90.Dryer, Matthew S. 1991. 'SVO Languages and the OV:VO Typology', Journal of Linguistics 27, 443-

482.Greenberg, Joseph H. 1963. 'Some Universals of Language with Special Reference to the Order of

Meaningful Elements', in J. Greenberg (ed.), Universals of Language, MIT Press,Cambridge, MA, pp. 73-113.

Hale, Kenneth. 1992. 'Basic Word Order in Two 'Free Word Order' Languages', in D. Payne (ed.),Pragmatics of Word Order Flexibility, John Benjamins, Amsterdam, pp. 63-82.

Hawkins, John A. 1994. A Performance Theory of Order and Constituency, Cambridge Studies inLinguistics, Vol. 73, Cambridge University Press, Cambridge.

Huang, C.-T. James. 1994. 'More on Chinese Word Order and Parametric Theory', in B. Lust, M.Suñer and J. Whitman (eds.), Syntactic Theory and First Language Acquisition: Cross-Linguistic Perspectives, Erlbaum, Hillsdale, NJ, pp. 15-35.

Hudson, Grover. 1972. 'Is Deep Structure Linear?' UCLA Papers in Syntax 2, 51-77.Kayne, Richard S. 1994. The Antisymmetry of Syntax, MIT Press, Cambridge, MA.Kiss, Katalin É. 1987. Configurationality in Hungarian, Studies in Natural Language and Linguistic

Theory, Reidel, Dordrecht.Lightfoot, David W. 1979. Principles of Diachronic Syntax, Cambridge Studies in Linguistics, Vol.

23, Cambridge University Press, Cambridge.Lust, Barbara, Margarita Suñer and John Whitman (eds.) 1994. Syntactic Theory and First Language

Acquisition: Cross-Linguistic Perspectives, Vol. 1: Heads, projections, and learnability,Erlbaum, Hillsdale, NJ.

Peterson, Thomas H. 1971. 'Multi-Ordered Base Structures in Generative Grammar', ChicagoLinguistic Society 7, 181-192.

Pierce, Amy. 1992. Language Acquisition and Syntactic Theory: A Comparative Analysis of French andEnglish Child Grammars, Kluwer, Dordrecht.

Pinker, Steven. 1984. Language Learnability and Language Development, Harvard University Press,Cambridge, MA.

Poeppel, David and Kenneth Wexler. 1993. 'The Full Competence Hypothesis of Clause Structurein Early German', Language 69, 1-33.

Roeper, Thomas and Jill De Villiers. 1994. 'Lexical Links in the Wh-Chain', in B. Lust, G. Hermonand J. Kornfilt (eds.), Syntactic Theory and First Language Acquisition: Cross-LinguisticPerspectives, Erlbaum, Hillsdale, NJ, pp. 357-390.

Sanders, Gerald A. 1970. 'Constraints on Constituent Ordering', Papers in Linguistics 2, 460-502.Staal, J. F. 1967. Word Order in Sanskrit and Universal Grammar, Foundations of Language

Supplementary Series, Vol. 5, Reidel, Dordrecht.Travis, Lisa. 1989. 'Parameters of Phrase Structure', in M. R. Baltin and A. S. Kroch (eds.),

Alternative Conceptions of Phrase Structure, University of Chicago Press, Chicago, pp. 263-279.

Valian, Virginia. 1991. 'Syntactic Subjects in the Early Speech of Italian and American Children',Cognition 40, 21-81.

Newmeyer Handout #2

14

5. PARSING AND BASIC WORD ORDER TYPOLOGY

(1) The Basic Word Order Paradox: Speakers rarely utter sentences with‘canonical’ constituent order (main clause declaratives in which allargument positions are filled by lexical NPs). But then why are typologiesbased on ‘canonical’ constituent order so useful?

Within the context of typological studies the term ‘basic order’ istypically identified with the order that occurs in stylistically neutral,independent, indicative clauses with full noun phrase (NP)participants … It must be remembered that the word ordercorrelations constituting the word order type… are established interms of the meta-theoretical concept of ‘basic order’. Therefore thedetermined correlations hold only for a limited subset of the actuallinearization patterns found in the language. (Siewierska 1988: 8, 25)

(2) Greenberg’s 6-way classification:VSO, SVO, SOV, VOS, OSV, OVS

LanguagesWord order Number %

SOV 180 45SVO 168 42VSO 37 9VOS 12 3OVS 5 1OSV 0 0

Total 402

Table 1Proportions of basic constituency orders

(Tomlin 1986: 22)

V-initial V-medial V-final42 30 64

Table 2Percent of languages of each type with explicit dependent (case) marking

(Siewierska and Bakker 1996)

(3) Examples of what Du Bois (1985) calls ‘Preferred Argument Structure’Cayuga (Iroquoian) — 1-2% of clauses contain 3 major constituents (Mithun

1987)Chamorro (Austronesian) — 10% of transitives have 2 lexical arguments

(Scancarelli 1985)Coos (Penutian) — 2-3% of clauses contain 3 major constituents (Mithun 1987)

Newmeyer Handout #2

15

French (Romance) — French preferred clause structure is [(COMP) clitic+Verb(X)]. Only 3% of clauses contain lexical subjects (Lambrecht 1987; thisconference)

German (Germanic) — even ditransitive verbs in spoken discourse tend to followPreferred Argument Structure (Schuetze-Coburn 1987)

Hebrew (Semitic) — 93% of transitive clauses lack a subject NP (Smith 1996)Huallaga Quechua (Andean) — in one corpus, only 8% of sentences contained

both a noun subject and a noun object (Weber 1989)Mam (Mayan) — 1% of clauses have 2 lexical arguments. (England 1988)Malay (Austronesian) — ‘Malay is thus similar to what Du Bois (1985) has

described for Sacapultec Maya: it has a “Preferred Argument Structure”’(Hopper 1988: 126)

Ngandi (Australian) — 2% of clauses contain 3 major constituents (Mithun 1987)‘O’odham = Papago (Uto-Aztecan) — only 9% of transitives have 2 overt

arguments (Payne 1992)Rama (Chibchan) — transitive clauses with 2 NPs are rare (Craig 1987)Sacapultec (Mayan) — in connected discourse, only 1.1% of clauses have 2 lexical

arguments. (Du Bois 1985; Du Bois 1987)Yagua (Peba-Yaguan) — in a corpus of 1516 clauses, only 3% contained both a

noun subject and a noun object (Payne 1990)

Even English, non-null-subject and considered rigidly SVO — a corpus of 20,794sentences (from telephone conversations) included only 5,975 (29%) thatwere SVO (Dick and Elman 2001)

(4) Two two-way typological parameters (Dryer 1991, Dryer 1997)a. SV vs. VSb. OV vs. VO

(5) Arguments for two two-way typological parameters (Dryer 1997)a. A number of languages are indeterminately VSO or VOS — the revised

typology classifies them identically (i. e., VO and VS).b. The fact that VSO and VOS have same the typological properties now

follows.c. The fact that VSO languages and VOS languages easily change from one

order to the other now follows.d. Cross-linguistically, clauses are rare with both an N subject and an N

objecte. Some languages can be classified by this typology, but are too

‘inconsistent’ with respect to the six-way typology.f. Some languages can be classified with respect to SV/VS , but not by

OV/VO, or vice-versa. Therefore, it is correct to separate out the two parameters.g. It is the OV/VO parameter that is central for typological predictions.h. The six-way typology does not distinguish between transitive subjects

and intransitive subjects, yet they often behave differently (e. g., Spanish istypically SVO, but VS).

Newmeyer Handout #2

16

Property V-final SVO V-initialPostpositional 96 14 09Relative-Noun 43 01 00Standard of comparison-Adjective 82 02 00Predicate-Copula 85 26 39Subordinate clause-Subordinator 70 06 06Noun-Plural word 100 24 13Adpositional phrase-Verb 90 01 00Manner Adverb-Verb 91 25 17Verb-Tense/aspect aux verb 94 21 13Verb-Negative auxiliary 88 13 00Genitive-Noun 89 59 28Sentence-Question particle 73 30 13

Wh-in situ 71 42 16Table 3

Percent of V-final, SVO, and V-initial languages manifesting particular properties(Dryer 1991)

(6) a. VSO languages are VS and VOb. SVO languages are SV and VO

(7) a. IP b. IP

NPsubj I' I'

I VP I VP

V' V NPsubj V'

V NPobj PP NPobj PP

P NP P NP

English — SVO Irish — VSO (McCloskey 1991)

V-initial V-medial V-final42 30 64

Table 2Percent of languages of each type with explicit dependent (case) marking

(Siewierska and Bakker 1996)

Newmeyer Handout #2

17

VSO VOSWithproperty

Withoppositeproperty

% withproperty

Withproperty

Withoppositeproperty

% withproperty

Prep 20 4 83 12 0 100NGen 24 3 89 11 3 79NRel 22 0 100 6 0 100ArtN 14 3 82 7 0 100NumN 19 5 79 8 1 89V-PP 20 0 100 8 0 100NegV 23 1 96 10 0 100AuxV 16 6 73 5 1 83Initial Q 12 7 63 3 0 100Initialwh

19 5 79 6 2 75

Table 4Word order characteristics of VSO and VOS languages

(Dryer 1997:77)

IP

(8) IP PPi

I' NPsubj P NP

I VP

V V'

NPobj PP

ei

A typical VOS language

IP(9)

NP VP

V NP

(10) a. subject - V b. V - object

Newmeyer Handout #2

18

(11) a. Marie voit Jean.b. Marie le voit.

Lemma structure plays a central role in the generation of surface structure.In particular, the main verb dictates what arguments have to be checkedin the message, and which grammatical functions will be assigned tothem. (Levelt 1989: 244)

(12) *Marie Jean voit.

(13) Who does Johni want to shave?(14) a. Himselfi

b. Himjc. Med. *Myselfe. *Himi

(15) a. Johni wants to shave himselfi.b. Johni wants to shave himj.c. Johni wants to shave me.d. * Johni wants to shave myself.e. * Johni wants to shave himi.

(16) a. Mary thinks that Sue has finally solved the problem.b. *Mary thinks Sue’s having finally solved the problem.c. *Mary thinks Sue to have finally solved the problem.

Notice the possible answers to (17)

(17) What does Mary think?

Only (18a) is possible, not (18b) or (18c):

(18) a. Sue has finally solved the problem.b. * Sue’s having finally solved the problem.c. * Sue to have finally solved the problem.

The internal structure of NP arguments has to be preserved as well, as (19a-c)show:

(19) a. Does Alice like the soprano?b. No, the tenor.c. *No, tenor.

REFERENCESCraig, Colette G. 1987. The Rama language: A text with grammatical notes. Journal of Chibchan

Studies, 5.Dick, Frederic and Jeffrey L. Elman. 2001. The frequency of major sentence types over discourse

levels: A corpus analysis. Newsletter of the Center for Research in Language, Universityof California, San Diego, 13.3-19.

Newmeyer Handout #2

19

Dryer, Matthew S. 1991. SVO languages and the OV:VO typology. Journal of Linguistics, 27.443-82.

Dryer, Matthew S. 1997. On the six-way word order typology. Studies in Language, 21.69-103.Du Bois, John. 1985. Competing motivations. Iconicity in syntax, ed. by John Haiman, 343-65.

Amsterdam: John Benjamins.Du Bois, John. 1987. The discourse basis of ergativity. Language, 63.805-55.England, Nora C. 1988. Mam voice. Passive and voice, ed. by Masayoshi Shibatani, 525-45.

Amsterdam: John Benjamins.Hopper, Paul J. 1988. Emergent grammar and the apriori grammar postulate. Linguistics in

context: Connecting observation and understanding, ed. by Deborah Tannen, 117-34.Norwood, NJ: Ablex.

Lambrecht, Knud. 1987. On the status of SVO sentences in French discourse. Coherence andgrounding in discourse, ed. by Russell Tomlin, 217-62. Amsterdam: John Benjamins.

Levelt, Willem J. M. 1989. Speaking: From intention to articulation. Cambridge, MA: MIT Press.Mccloskey, James. 1991. Clause structure, ellipsis, and proper government in Irish. Lingua,

85.259-302.Mithun, Marianne. 1987. Is basic word order universal? Coherence and grounding in discourse,

ed. by Russ Tomlin, 281-328. Amsterdam: John Benjamins.Morgan, Jerry L. 1973. Sentence fragments and the notion 'sentence'. Issues in linguistics: Papers

in honor of Henry and Renée Kahane, ed. by Braj B. Kachru, Robert B. Lees, YakovMalkiel, Angelina Pietrangeli and Sol Saporta, 719-51. Urbana: University of IllinoisPress.

Payne, Doris L. 1990. The pragmatics of word order: Typological dimensions of verb-initiallanguages. Berlin: Mouton de Gruyter.

Payne, Doris L. 1992. Nonidentifiable information and pragmatic order rules in 'O'odham.Pragmatics of word order flexibility, ed. by Doris L. Payne, 137-66. Amsterdam: JohnBenjamins.

Pope, Emily. 1971. Answers to yes-no questions. Linguistic Inquiry, 2.69-82.Scancarelli, Janine S. 1985. Referential strategies in Chamorro narratives: Preferred clause

structure and ergativity. Studies in Language, 9.335-62.Schuetze-Coburn, Stephan. 1987. Topic management and the lexicon: A discourse profile of three-

argument verbs in German: UCLA MA thesis.Siewierska, Anna. 1988. Word order rules. London: Croom Helm.Siewierska, Anna and Dik Bakker. 1996. The distribution of subject and object agreement and

word order type. Studies in Language, 20.115-61.Smith, Wendy. 1996. Spoken narrative and preferred clause structure: Evidence from Modern

Hebrew discourse. Studies in Language, 20.163-89.Tomlin, Russell S. 1986. Basic word order: Functional principles. London: Croom Helm.Weber, David J. 1989. A grammar of Huallaga (Huánaco) Quechua. Berkeley: University of

California Press.

Newmeyer Handout #2

20

6. TYPOLOGY, UNIVERSAL GRAMMAR, AND LANGUAGE EVOLUTION

TWO ASSUMPTIONS ABOUT SOV VIS-A-VIS SVO ORDER:

(1) SOV order predominates among the languages spoken today.Afr Eura A-NG NAm SAm Total

SOV 22 26 19 26 18 111SVO 21 19 6 6 5 57VSO 5 3 0 12 2 22

Breakdown of genera in terms of basic word order, by area (Dryer 1989)

(2) The historical change OV > VO is both more common than the change VO> OV (Li 1977) and more ‘natural’ (Aske 1998; Vennemann 1973).

(3) Many have argued that OV > VO is more ‘natural’ than VO > OVa. Vennemann (1973) points to the loss of case endings in SOV

languages through phonological attrition. In such an eventuality, the shift of theverb to the middle serves to demarcate more saliently the subject from the object.

b. Hawkins (1994) notes that SOV languages with ‘heavy’ objectscreate considerable parsing difficulty. Processing efficiency is maximallyimproved by preposing the object to initial position, creating OSV order.Efficiency is also increased (though not as greatly) by postposing the object,creating SVO order. Thus there is a ready parsing-based mechanism for ‘leakage’from SOV to SVO. It is never advantageous from the point of view of parsing foran SVO language to prepose a heavy object to a position between the subject andthe object. And hence, no there is nothing based in language processing thatwould create ‘leakage’ from SVO to SOV.

c. Aske (1998) points to discourse-based pressure for the developmentof a focus position after V in OV languages, leading ultimately to VO order.However, there is no corresponding mechanism for the loss of an NP after V inVO languages. Thus he posits a long-term drift to VO.

(4) SOV languages are more likely to have alternate orderings of S,V, and Othan do languages with other basic orderings:

VOS VSO SOV SVO

very common VSO VOS OSV ------ SVO

common ------ SVO SVO -------

Alternate orderings for basic orders (Steele 1978)

(5) INTERIM CONCLUSION: SOV order was once much more typologicallypredominant than it is now.(6) Protolanguage had thematic structure (Bickerton 1990).

Newmeyer Handout #2

21

(7) Protolanguage lacked quantificational structure.

(8) A NEW OBSERVATION ABOUT VO vs. OV TYPOLOGY:a. VO languages are ‘good at’ representing quantification directly, but

‘bad at’ representing thematic structure directly.b. OV languages are ‘good at’ representing thematic structure

directly, but ‘bad at’ representing quantification directly.

English, a typical SVO language, the verb find, say, can take an agent, a theme, ora locative as a subject:

(9) a. Mary found what she was looking for.b. My old sofa found a home in the departmental lounge.c. Noon found Bill eating lunch in the Student Union.

(10) English (VO) vs. German (OV) (Hawkins 1995):a. Subject-to-Subject Raising

E: The noise ceased to get on his nerves.G: ?Der Lärm hörte auf, ihn aufzuregen.

b. Tough MovementE: Literature is boring to study.G: *Die Literatur ist langweilig zu studieren.

c. Subject-to-Object RaisingE: I believe the farmer to have killed the cow.G: *Ich glaube den Bauern, die Kuh geschlachtet zu haben.

d. Out-of-clause Wh-MovementE: What did you assume that we would not bring?G: *Was hast du angenommen, dass wir nicht mitbringen würden?

e. Semantic diversity of grammatical relationsE: My guitar broke a string.G: *Meine Gitarre hat eine Seile zerrissen.

f. Breadth of subcategorization possibilitiesE: This door will open (new possibilities).G: Diese Tür wird sich öffnen / *neue Möglichkeiten öffnen.

(11) Universal 41 (Greenberg 1963: 113)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.

(12) In the sample of 237 languages presented in Siewierska and Bakker (1994),64% of SOV languages have explicit case, but only 30% and 42% of SVO and VSOlanguages respectively.

An SVO underlying structure like (13a) can underlie (13b) and (13c), where thesurface position of who uniquely identifies the former as an indirect question andthe latter as a direct question:(13) a. He was wondering [you saw who]

b. He was wondering who you sawc. Who was he wondering that you saw?

Newmeyer Handout #2

22

But with no movement of the wh-operator, an SOV underlying structure like(14a) represents the order of elements both in the indirect and the direct question:(14) he [you saw who] were wondering =

a. Who is he wondering that you saw? orb. He was wondering who you saw.

(15) breakFrame 1. NP — V — NP John broke my guitar

[Agent] [Patient]2. NP — V My guitar broke

[Patient]3. NP — V — PP A string broke on my guitar

[Patient] [Locative]4. NP — V — NP My guitar broke a string

[Locative] [Patient]5. NP — V — NP My guitar broke a world record

[Instrument] [Patient]

(16) believeFrame 1. NP — V — NP I believe the farmer

[Agent] [Dative]2. NP — V — NP I believe this report

[Agent] [Patient]3. NP — V — S' I believe that the farmer killed

the cow[Agent] [Object of belief]

4. NP — V — NP — VP' I believe the farmer to have killed the cow

[Agent] [Agent]

(17) CONCLUSION: THE EARLIEST HUMAN LANGUAGE WAS OV

(18) 3 BROAD SCENARIOS FOR THE ORIGIN OF UG CONSTRAINTS:a. Big Bang theories — the evolutionary ‘event’ responsible for

human language was responsible for UG constraints (Chomsky 1988; Chomsky1991; Bickerton 1990)

b. Genetic assimilation theories — UG constraints resulted from thenativization of parsing (and other performance?) principles via geneticassimilation (the Baldwin Effect) (Berwick and Weinberg 1984; Kirby andHurford 1997)

c. Constraints as epiphenomena theories — there is no evolutionaryissue per se, since ‘UG constraints’ are in reality epiphenomenal by-products ofmechanisms not particular to grammar (Deane 1992; Kluender 1992)

(19) [[[ei osiete-ita] seito ga] rakudaisita] senseii ‘teaching-was student flunked teacher’

the teacher who the students that (he) was teaching flunked

Newmeyer Handout #2

23

(20) John-wa Mary-ga nani-o katta kadooka siritagatte iru noJohn wants to know whether Mary bought what

(21) NP

S NP

senseiiNP VP

rakudaisitaS NP

seito-gaNP VP

ei osiete-ita

(22) cakicasini / pro-casini -GEN mother-NOM Johni betrayed / hithimself’s mother John betrayed

(23) Ranked constraints in the binding of Korean Long Distance Anaphors(Moon 1995):a. Thematic Hierarchy Constraint (LDA must be bound by a thematically higherNP)b. Larger Domain Preference Constraint (Given potential antecedents for LDA indifferent domains, the more distant the domain, the stronger the preference)c. Subject-Orientation Constraint (LDA must be bound by a subject NP)d. C-Command Constraint (LDA must be bound by a c-commanding NP)e. Discourse Binding Constraint (LDA must be bound by a prominent discourseNP if no sentential antecedent is available)

(24) *John-wa Mary-ga naze sore-o katta kadooka siritagatte iru no?‘John wants to know [whether [ Mary bought it why]]?’

REFERENCES

Aske, Jon. 1998. Basque Word Order and Disorder: Principles, Variation, and Prospects, JohnBenjamins, Amsterdam.

Berwick, Robert C. and Amy Weinberg. 1984. The Grammatical Basis of Linguistic Performance, MITPress, Cambridge, MA.

Bickerton, Derek. 1990. Language and Species, University of Chicago Press, Chicago.Chomsky, Noam. 1988. Language and Problems of Knowledge: The Managua Lectures, Current Studies

in Linguistics, Vol. 16, MIT Press, Cambridge, MA.Chomsky, Noam. 1991. 'Linguistics and Cognitive Science: Problems and Mysteries', in A. Kasher

(ed.), The Chomskyan Turn: Generative Linguistics, Philosophy, Mathematics, and Psychology,Blackwell, Oxford, pp. 26-55.

Deane, Paul D. 1992. Grammar in Mind and Brain: Explorations in Cognitive Syntax, CognitiveLinguistics Research, Vol. 2, Mouton de Gruyter, The Hague.

Newmeyer Handout #2

24

Dryer, Matthew S. 1989. 'Large Linguistic Areas and Language Sampling', Studies in Language 13,257-292.

Hawkins, John A. 1995. 'Argument-Predicate Structure in Grammar and Performance: AComparison of English and German', in I. Rauch and G. F. Carr (eds.), Insights inGermanic Linguistics I, Mouton de Gruyter, Berlin, pp. 127-144.

Kirby, Simon and James Hurford. 1997. 'Learning, Culture, and Evolution in the Origin ofLinguistic Constraints', in P. Husbands and H. Inman (eds.), Proceedings of the FourthEuropean Conference on Artificial Life, MIT Press, Cambridge, MA, pp. 493-502.

Kluender, Robert. 1992. 'Deriving Island Constraints from Principles of Predication', in H.Goodluck and M. Rochemont (eds.), Island Constraints: Theory, Acquisition, and Processing,Kluwer, Dordrecht, pp. 223-258.

Li, Charles N. (ed.). 1977. Mechanisms of Syntactic Change, University of Texas Press, Austin.Moon, Seung Chul. 1995. An Optimality Approach to Long Distance Anaphors, Ph. D. dissertation,

University of Washington.Newmeyer, Frederick J. 2000. 'On Reconstructing 'Proto-World' Word Order', in C. Knight, J.

Hurford and M. Studdert-Kennedy (eds.), The Emergence of Language, CambridgeUniversity Press, Cambridge, pp. 372-388.

Siewierska, Anna (ed.). 1994. Constituent Order Working Paper 6, ESF Eurotype Project, Strasbourg.Steele, Susan. 1978. 'Word Order Variation: A Typological Study', in J. H. Greenberg, C. A.

Ferguson and E. A. Moravcsik (eds.), Universals of Human Language, Stanford UniversityPress, Stanford, CA, pp. 585-623.

Vennemann, Theo. 1973. 'Explanation in Syntax', in J. Kimball (ed.), Syntax and Semantics, Vol 2,Seminar Press, New York, pp. 1-50.