Word Order

Word Order

            The positioning of the subject (S), verb (V), and object (O) in informationally neutral clause (without topic and focus ) id one of the major fifferences among languages. Permutations of the three clause inredients can give six possible word orders. However, not all six are equally attested in the world languanges. Typological generalization show that in the vast majority of human languanges, the basic word order is either SOV (about 45% in the world Atlas of Language Structure) or SVO (about 35.4 %) while orders like VSO (9%) or OSV (0.5%) are much less frequent or extremely rare.
To date, most L2 research has focused on acquisition of one or the order most prevalent word order. How does a learner acquire basic SVO when their native languange axbihits or vive versa ?  the theoritical conceptualization of this parameter has changed over the years. At the time when researchers were interested on OV/VO word order acquisition, the prevelant analysis was that be ordering of the verbal head and the object in the VP wasa parameter provided by UG and fixed on the basis of ample linguistic experience.

Head-initial VP (english,romance) and Head-initial VP (Japanese, German )

            In the 1990’s, an influential proposal by Kayne (1994) suggested that SVO was the universal underlying word order, and all other attested orders were a  product of constituent movement. In the 2000s, the analysis shifted again to reflect a minimalist UG cleansed of all parameter .it is now widely accepted that VP is built through the operation of Merge, satisfyimg the selection features on the verb ( e.g the verb eat is transitive and hence need to merge with two arguments, an Agent and a Theme ). Once the VP is built,  the linearization of the string, that is, where each constituent is pronounced in the surface string, is calculated at the interface between syntaxand the sensory-motor system, and depands on adiacritic on the verb. Essentially, the lexical infoemation of the verb in English contains the information the verb goes before the object, while the lexical information of the Japanese verb contains the opposite information, that the verb follows the object.

            Let us consider the experimental evidance now, with respact to the word order in VP property (OV vs. VO). A common assumption of The Full Transfer Full Access Hypothesis and The Minimal Trees  Hypothesis is that learners initially transfer their native word order in the languange they are acquiring. For learners od  German (SOV) whose native languange  is a Romance Languange (SVO), the expectation is that they will initially tranfer SVO and will place the verb ( boyh finite and non-finite forms) before the object. This assumption is suported by findings in Meisel. Clashen, and Pieneman (1981), hwhere verbs initially appeard in second position. Supporting findings also come from  Vainikka and Young-Scholten’s work : Korean and Turkish learners of German are observed to prfer verb-final constructions initially, while Italian and Spamish learners tend to use  left-headed VPs. The fact that learner use their native VP word appears to be empirically well established.

            How about acquisition of word order more generally speking  ?  beyond the order of constituents in the vp, surface word order is regulated by uninterpretable features. One of the most important of such features, known as the EPP features of historical reasons, control wheather subjects have to move to the TP projection, or whether another constituent or a null subject can check the features. Also regulated by uninterpretable features is the movement of nominal and other constituent called “scrambling (think : scrambled eggs). In the languages where scrambling is licit, such constiuents optionally, sometimes to such feature acqusition next.

            (2011), a large study investingating knowladge of word order and its interaction with case marking and specificity in the interlanguange of Greek native speaker, beginning and intermediate learners of Turkish. Greek exhibits SVO word order  while Turkish is an SOV languange so learners have to establish a new word order in their second languange. The battery of tests in this study included a cloze test, an online grammaticality more accurate on verbal inflection (agreement) than on nominal inflection (case endings), as established in the cloze test. Accurancy percentages for the beginner group were 42 % versus 21%, respectively. At the time, increased proficiency (95% for verbal inflection, 49% for case in the most proficient group ).

            In the grammaticality judgement task, which utilized only simple transitive sentence, learners viewed each word centered on the ecreen for 2 seconds. A “?” at the end of sentence prompted them to record whether they findings the sentence acceptable. Response times were measured. The study findings attest to the target SOV word order of Turkish being acquired even by the lowest proficientcy group. Howewver, om non-canonical world arders, learner did not do so well. Look at the fllowing examples.

1.      Newspaper-PL-ACC child read-PAST  ‘the newspapers were read by some child’ O_ACCSV

2.      Newspaper child read-PAST ‘newspaper were read by some child’ O _ABS S V

3.      I think that peter already tried to repair the car

4.      I think taht peter already tried to repair the car.

            Knowladge of the whole paradigm (allowed and disallowed scrembled sentences ) depends on uninterpretable features beyond those responsible for topic alization. In addition, Japanese learner would have an advantages over english native speakres in learning these construction, since Japanese also allowes to them from their native language, while it would be new for the English and Japanee learnes robustly repeated the evaluations of the native German speakers. He also uncovered native languanges influence. Hopp  made the case that knowladge of the allowed and diasallowd sentence could not have come trough classroom instructionor observation of the native input. Since complete knowladge of this paradigm involves rejection of unacceptable constructions that cannot come from astute observation of the input and pattern-noticing, Hopp argued that is a proverty of the Stimulus learning situation. The input simply does not contain unaccetable sentence of the sort of (5). The input does not contain unacceptable sentence in (4), and both are pretty complex. Still., advanced learners distinguish between them reliably.

           

            So,the relative merits of the two opposing postion, eith respect to word order phenomena. Proponents of representational deficits can take heart from findings such as Papadopoulou et al.’s (2011), because they document lack of successful acqusition of non-canonical word orders. For full functional representation proponents, however, findings such as Hopp’s (2005) provide a clinching argument, acquisition of initerpretable features cannot be impossible, if is crucial that successful acquisition involves knowladge of ungrammaticality, which can not be due to noticing patterns or imition. However, pattern noticing and imitation are the only ways in which deficit accounts can explain learner success with uninterpretable features acquisition.