View sample Word Order Research Paper. Browse other research paper examples and check the list of research paper topics for more inspiration. If you need a religion research paper written according to all the academic standards, you can always turn to our experienced writers for help. This is how your paper can get an A! Feel free to contact our custom writing service for professional assistance. We offer high-quality assignments for reasonable rates.
As is well known, words in clauses and phrases are not strung together in a random fashion but are ordered in a communicatively signiﬁcant way. Thus, for example, John hit Mary does not mean the same thing as Mary hit John. Moreover, the word orderings that are communicatively signiﬁcant are not cross-linguistically uniform. Thus, whereas in English neither of the events described in the above two italicized clauses could be expressed with the verb in initial position, i.e., as Hit Mary John, this is precisely the order of words that would be used in Samoan or Berber. The nature of the word order patterns found within and across languages and the factors underlying their occurrence and distribution has been a topic of intensive study from a number of different perspectives. This article provides an overview of functional—typological research on word order concentrating on three issues, namely, the classiﬁcation of languages in terms of their basic word order, the determination of the degree of language-internal word order variation, and the attempts to formulate a unifying principle of order capable of accounting for both cross-linguistic and language internal word order patterns.
1. Basic Word Order Typology
1.1 The Clause
Ever since Greenberg (1963) it has been the custom to classify the clausal word order of a language in terms of the location of the verb and its arguments relative to each other in transitive clauses (as opposed to intransitive or ditransitive ones) into one of the following six types: SOV, SVO, VSO, VOS, OVS, OSV, where S stands for subject, O for object, and V for verb. The transitive clauses which have served as the basis for word order classiﬁcation are active, positive, declarative clauses with two nominal (not pronominal) participants, a deﬁnite human agent, and a deﬁnite nonhuman patient such as The man killed the snake, or The girl extinguished the ﬁre. Greenberg took the order obtaining in such clauses to be the basic order of a language. For languages exhibiting more than one arrangement of the S, O, and V in the relevant type of clauses, the basic order was identiﬁed as the statistically dominant order.
Nowadays it is generally recognized that not all languages fall into one of the six word-order types. First of all, there are languages which simply lack transitive clauses with two overt NP participants, such as Puget Salish. Second, in many other languages, though clauses with two overt nominal participants do exist, they are quite rare, as both of the arguments must, or tend to, be expressed by pronominal affixes on the verb or by clitics elsewhere in the clause. For example, in her analysis of narrative texts in the South American language, Yagua, Payne (1990) found only 45 transitive clauses with two overt nominal participants, which constitutes less than 10 percent of the transitive clauses (463) and only 3 percent of all the clauses (1516) in the texts considered. In such languages, it may not always be possible to discern a clearly dominant order among the verb and its nominal arguments. Dryer (1997) suggests that the basic order should be at least twice as frequent as the order or orders with which it contrasts. In terms of such a criterion, none of the transitive orders in Yagua qualify as basic, as the frequencies noted by Payne were: VSO (19), SVO (15), and OVS (11). The third type of languages which may fall outside the six-way typology are languages in which, while transitive clauses with two overt nominal participants are quite frequent, no single order may unequivocally be dominant, as the ordering of constituents may be determined not by their grammatical relations but rather by their pragmatic status and discourse role.
In view of the above, several modiﬁcations of Greenberg’s typology are currently in use. One involves using double classiﬁcations such as VSO/VOS or SVO/VSO. Another involves augmenting the six-way typology by a seventh type, namely, no basic order. A third modiﬁcation consists in reanalyzing the six orders in terms of a three-way opposition of verb position into verb-initial, verb-medial, and verbﬁnal. The most radical modiﬁcation is that suggested by Dryer (1997), who argues for replacing the six-way typology by two separate two-way typologies of SV vs. VS and OV vs. VO, which together deﬁne four types: VS & VO, SV & VO, SV & OV, and VS & OV. While this typology also concerns the order of nominals in transitive clauses, it differs from all of the above in being based on the presence of only one nominal participant, as opposed to two. Accordingly, it is more widely applicable and easier to apply than the six-way typology.
Despite the problems posed in classifying some languages in terms of their basic transitive order, the cross-linguistic investigations conducted since the 1960s identify the same basic word order preferences consistently; the most common basic order is SOV (+ / – 45 percent), followed closely by SVO ( + / – 35 percent), and then VSO ( + / -10 percent). The three object-before-subject orders are all highly uncommon as basic orders, with VOS occurring slightly more frequently than OVS, and OSV being extremely rare.
Turning to the basic order in other than transitive clauses, with few exceptions (e.g., Salinan, Woods Cree, Makushi), the basic order of the sole argument of the intransitive verb is the same as that of the S argument of a transitive clause. Consequently, the classiﬁcation of a language’s transitive order also covers that of intransitive clauses. This also holds to a large extent for ditransitive clauses, i.e., clauses featuring, in addition to an agent and patient, a recipient (e.g., Mother gave the baby a toy.) Again with relatively few exceptions (e.g., Nadeb, Trumai, Yoruba) recipients tend to be placed on the same side of the verb as the O in transitive clauses, and typically occur contiguous to the patient. According to Primus (1997), in the basic order recipients are never separated from the patient by the agentive argument.
1.2 The Phrase
Order within phrases is characterized typically in terms of a two-way typology originating in the dependency tradition and the work of the nineteenth-century German typologists. This two-way typology is based on the division of the constituents of phrases into head / modiﬁer pairs such as: verb / object, verb / adverbial, auxiliary / main verb, adposition / noun phrase, noun / adjective, noun / genitive, noun / relative clause, etc. (where the ﬁrst member of the pair is the head and the second the modiﬁer). As reﬂected in Vennemann’s (1972) natural serialization principle and its analogs, languages are assumed to show a preference for serializing constituents with reference to either the head > modiﬁer or the modiﬁer > head schema. Those that exhibit basic transitive VO order are expected to display head > modiﬁer order, and those that have basic OV order, modiﬁer > head patterns. Languages that conform to these expectations are termed consistent. Consistency, however, is viewed currently not as the cross-linguistic norm, but rather as an ideal to which languages approximate to various degrees. And indeed very few languages are absolutely consistent in ordering all their modiﬁers either to the left or the right of the head. Thus, for example, of the 142 languages in Greenberg’s Appendix II, only 68 (48 percent) exhibit consistent ordering in relation to just four head / modiﬁer pairs, i.e., the verb and object, adposition and noun phrase, noun and genitive, and noun and adjective. Needless to say, the number of consistent languages drops once further head / modiﬁer pairs are considered.
Several reﬁnements to the expectation of consistent ordering have been suggested. One of these is the principle of cross-category harmony (PCCH) formulated by Hawkins (1983). The PCCH states that languages display a preference to generalize the order obtaining in one head / modiﬁer category to that of other categories. Departures from this preference in one head / modiﬁer category are in turn likely to induce comparable departures in other head / modiﬁer pairs. Thus, for example, Hawkins notes that the continuum in verb position, verb-initial, verb-medial, and verb- ﬁnal is reﬂected in the position of the noun in relation to its modiﬁers, VSO languages being more consistently noun-initial than SVO ones, and non-rigid SOV languages (SOV languages which allow for the placing of some modiﬁers to the right of the verb) less consistently noun-ﬁnal than rigid SOV ones.
Another reﬁnement to the preference for consistent ordering has been introduced by Dryer (1992) who argues that this preference actually involves not head/modiﬁer pairs but rather lexical and phrasal categories. Dryer suggests that languages tend to be either right-branching, in which phrasal categories follow non-phrasal categories, or left-branching, in which phrasal categories precede non-phrasal categories. This is argued to hold irrespective of which is considered to be the head and which the modiﬁer. For non-branching categories, no word-order predictions are made. Thus, according to Dryer’s branching direction theory, whereas, for example, a genitive phrase (e.g., of John) and relative clause (whom I met) should occur on the same side of the noun, a demonstrative phrase need not, as it is not a branching category; nor need an adjective pattern as the genitive phrase or relative clause, since it is not a fully recursive phrasal category. While Dryer’s word order pre- dictions receive strong cross-linguistic support, their correctness is heavily dependent on speciﬁc theories of constituent structure which are not uncontroversial.
The notion of consistent ordering in its various guises relates to the order of head / modiﬁer pairs and says nothing about the ordering of modiﬁers relative to each other. This latter aspect of phrasal order is still under investigation. One subset of modiﬁers whose order has received detailed attention is that of the demonstrative, numeral, and adjective within simple NPs. Rijkhoff (1997) accounts for the cross-linguistic linearizations of these three modiﬁers with reference to three principles, namely, the principle of domain integrity, the principle of head proximity, and the principle of scope. All three principles are elaborations of what is known as Behagel’s (1932, p. 4) ﬁrst law, which is ‘what belongs together semantically is also placed together syntactically.’ Of the 24 logically possible linearizations of the demonstrative, numeral, adjective, and noun, the three principles deﬁne the following eight word-order patterns as potential basic orders.
- DemNumAdjN (b) AdjNNumDem
Rijkhoff’s cross-linguistic data reveal that all these orders, and only these, are attested as basic orders.
2. Word Order Variation
Languages are classiﬁed not only in regard to their basic order, but also with respect to the amount of word order variation that they display. However, though labels such as rigid or ﬁxed-order language, or free or ﬂexible-order language are commonplace, there is little consensus on what they actually mean. This is not altogether surprising as it is far from clear how word-order variation could be measured.
The major problem in devising a measure of wordorder variation is the sheer number of potential variables involved. An alternative order may be conditioned by a wide range of factors—syntactic, semantic, and pragmatic—relating to the nature of the text and/or genre (e.g., written vs. spoken, narrative vs. expository vs. procedural), the properties of the clause (e.g., its illocutionary force or main vs. subordinate status), the features of the verb (e.g., its voice, tense, or aspect) and the characteristics of its arguments and adjuncts such as referentiality, deﬁniteness, givenness, humanness, animacy, semantic role, and morphological and syntactic complexity. Needless to say, there is no obvious reason why changes in order attributable to any one factor or set of factors should be viewed as being more signiﬁcant than others.
The above notwithstanding, some idea of the wordorder ﬂexibility of a language can be gathered by reducing the number of variables and considering the number of word-order permutations in clauses or head/modiﬁer pairs exhibiting the same characteristics as those used for determining the basic order. In terms of such a measure, English emerges as a fairly rigid word order language at the clausal level, as the only variation on its basic SVO order that it permits is OSV. Polish, on the other hand, is a highly ﬂexible word order language, since in addition to its basic SVO order it also has all the other ﬁve permutations of the S, O, and V. Applying the above measure of wordorder ﬂexibility to a sample of 171 1anguages, Siewierska (1997) found no variation on the basic transitive order in 22 percent of the languages, only one variant in 35 percent, two variants in 17 percent, three variants in 12 percent, four variants in 6 percent, and ﬁve variations in 8 percent. Thus languages with no or little word-order ﬂexibility appear to be considerably more common than those with highly ﬂexible order.
Some generalizations can also be made with respect to the type of word order variations that languages are most and least likely to exhibit, namely (a) verb initial languages are least likely to exhibit verb ﬁnal orders, and verb ﬁnal languages, verb initial orders, (b) in languages of all word order types, SVO is a more likely variable order than OVS. Thus, an SOV language with fairly inﬂexible order may be expected to have an OSV variant, or possibly a SVO one, but not a VSO or OVS variant. And a VSO language may be expected to exhibit SOV order only if it also has VOS, SVO, and OVS.
The degree of word order ﬂexibility of a language with respect to the ordering of the verb and its arguments is often seen to be tied to the existence of agreement and case marking. The argument is that, if the subject is distinguished morphologically from the object by case and/or agreement marking, word order is ‘freed’ from performing this function and can thus be used for other communicative ends. This, however, is not necessarily so. Cross-linguistic data cited by Siewierska (1997) reveal that, contrary to what is often suggested, it is not the presence of case and agreement marking that has an affect on word-order ﬂexibility but rather their absence. Rigid order is associated strongly with lack of agreement and to a lesser extent with lack of case marking, but neither the presence of case nor agreement marking correlates with ﬂexible order.
The investigation of word-order variation carried out to date suggests that languages exhibit more variation in the ordering of constituents at the level of the clause than at the level of the phrase. Signiﬁcant differences in word-order ﬂexibility have also been observed in regard to speciﬁc head/modiﬁer pairs. Bakker’s (1997) investigations of the possibility of alternative orderings of the noun and a single modiﬁer among the languages of Europe suggests that the modiﬁers of the noun can be placed in a hierarchy of decreasing likelihood of exhibiting an alternative order to that of their basic order, namely: adjective > genitive > relative clause > numeral > demonstrative. Whether this will be. conﬁrmed globally remains to be determined.
3. Discourse-Pragmatic Vs. Processing, Explanations Of Order
Both the cross-linguistic preferences in regard to the distribution of basic orders as well as the type of variations on the basic order that languages tend to favor have typically been accounted for in discourse-pragmatic terms. Until recently, the dominant principle assumed to underlie linearization has been the Topic > Comment principle, which deﬁnes a preference for placing constituents encoding more topical, predictable, given information before those encoding less topical, unpredictable, new information. The universal applicability of this principle has been called into question by the existence of languages which appear to favor placing the comment not after, but before, the topic. Givon (1988) has sought to reconcile the existence of two opposing principles of order by formulating the principle of Task Urgency, which is: ‘Attend ﬁrst to the most urgent task,’ where a communicative task is identiﬁed as more urgent if the information to be communicated is either less predictable or more important. Givon suggests that when the topic of discourse is a highly predictable referent, the most urgent task is to present the comment, and when the topic corresponds to a less predictable or unpredictable discourse referent, the most urgent task is to state the topic. Thus the principle of Task Urgency predicts a cross-linguistic preference for placing either less predictable > more predictable and/or more important > less important information.
An alternative unifying principle of order has been posited recently by Hawkins (1994), who sees order as being determined primarily by processing considerations and not discourse-pragmatic ones. Building on the long-recognized tendency to place short constituents before longer ones, Hawkins hypothesizes that the human parser displays a preference for linearization patterns which allow the quickest recognition of immediate constituents of syntactic groupings. This is captured in the Early Immediate Constituent recognition principle (EIC), which is: ‘The human parser prefers to maximize the left-to-right IC (immediate constituent)-to-word ratios of the phrasal nodes that it constructs’. The EIC predicts that information about immediate constituency should be placed in the earlier rather than the later part of a string. For relatively heavy categories, such as sentential and NP complements, this prediction entails a preference for short > long linearization in right-branching ( VO) languages and for long > short in left-branching (OV) languages. But in the case of single-word categories, an overall preference for placing the light category to the far left is deﬁned.
Strong cross-linguistic evidence for the EIC has been cited from several types of data: real-time psycholinguistic experiments of alternative orderings, of ICs, language-internal text frequency counts, native speaker acceptability judgments, and the cross-linguistic distribution of basic orders. These data show that not only is there an unequivocal preference for orderings deﬁned as optimal by the EIC, but also that there is a progressive decline in the frequency of nonoptimal orders directly proportional to the degree of their non-optimality deﬁned by the EIC. Thus, for example, while the EIC deﬁnes the presence of both prepositions in OV languages, and of postpositions in VO as non-optimal, the EIC score for the former, e.g., for with the friends played is considerably lower (20 percent) than for the later, e.g., played the friends with (65 percent). And though both orders are cross-linguistically uncommon, in Dryer’s (1992) data, OV languages with prepositions are three times less common than VO languages with postpositions (1.5 percent vs. 4.6 percent).
Whether an EIC account of word order is indeed superior to one based on discourse-pragmatic principles, as Hawkins argues, remains to be seen.
- Bakker D 1997 Flexibility and consistency in word order patterns in the languages of Europe. In: Siewierska A (ed.) Constituent Order in the Languages of Mouton de Gruyter, Berlin
- Behaghel O 1932 Deutsche Syntax: eine geschichtliche Darstellung. Band IV: Worststellung-Periodenbau. Carl Winter, Heidelberg, Germany
- Dryer M 1992 The Greenbergian word order Language 63: 81–138
- Dryer M 1997 On the six-way word order Studies in Language 21: 69–103
- Givon T 1988 The pragmatics of word order: Predictability, importance and attention. In: Hammond M, Moravcsik E A, Wirth J R (eds.) Studies in Syntactic Typology. John Benjamins, Amsterdam, The Netherlands
- Greenberg J A 1963 Some universals of grammar with particular reference to the order of meaningful In: Greenberg J (ed.) Universals of Language. MIT Press, Cambridge, MA
- Hawkins J A 1983 Word Order Academic Press, New York
- Hawkins J A 1994 A Performance Theory of Order and Constituency. Cambridge University Press, Cambridge, UK
- Payne D 1990 The Pragmatics of Word Typological Dimensions of Verb Initial Languages. Mouton de Gruyter, Berlin
- Primus B 1997 The relative order of the patient and recipient in the languages of Europe. In: Siewierska A (ed.) Constituent Order in the Languages of Mouton de Gruyter, Berlin
- Rijkhoff J 1997 Order in the noun phrase of the languages of Europe. In: Siewierska A (ed.) Constituent Order in the Languages of Europe. Mouton de Gruyter, Berlin
- Siewierska A 1997 Variations in major constituent order: A global and a European perspective. In: Siewierska A (ed.) Constituent Order in the Languages of Mouton de Gruyter, Berlin
- Vennemann T 1972 Analogy in generative grammar, the origin of word In: Heilmann L (ed.) Proceedings of the Eleventh International Congress of Linguistics, Vol. 2. Il Mulino, Bologna, Italy