Cross-linguistic evidence for processing strategies
المؤلف:
Paul Warren
المصدر:
Introducing Psycholinguistics
الجزء والصفحة:
P171
2025-11-10
60
Cross-linguistic evidence for processing strategies
The examples discussed above have all been from English, chiefly because English is the language of this book, and it is therefore assumed that the English examples will be most widely accessible to readers.
Much of the early research on parsing strategies did in fact focus on English, and assumed that strategies were applied equally regardless of the availability of other information. MacWhinney, Bates and Kliegl (1984), however, ran a study of English, German and Italian that showed that there is interaction between strategies and that speakers will use whatever information they can to understand a sentence. MacWhinney et al. asked native speakers of these three languages to carry out a simple task with short sentences in their native language. The task was to identify the agent the entity that instigates the action shown by the verb in simple sentences such as those in (10.53)–(10.55). These sentences were especially constructed so that the authors could examine participants’ use of cues based on word order, noun–verb agreement such as the grammatical marking of subject and object through inflections in certain languages and animacy of the nouns.
The authors found that native speakers of English relied primarily on word order, native speakers of German on animacy, and to a lesser extent on noun–verb agreement, and native speakers of Italian on noun–verb agreement, and to a lesser extent on animacy. Clearly a key factor in learning a strategy for processing a language is the reliability of the cues employed by that strategy. In English, word order is relatively fixed, and noun-verb agreement is absent apart from in the pronominal system (he/him). Therefore, it is not surprising to find native English speakers relying mainly on word order. Studies like this also show how just looking at data from English gives an incomplete view of language processing.
Further interesting work on sentence processing has compared how a particular strategy applies across different languages for very similar types of syntactic ambiguity. Particularly relevant here is the case of relative clause attachment ambiguity. Ambiguity can arise if there is more than one preceding noun that could be modified by the relative clause, as in (10.56). The ambiguity concerns who in fact had had the accident – the colonel or his daughter?
Following the parsing strategy of Late Closure, the relative clause should be attached to the second noun phrase in the object construction, and so will be taken to modify colonel. Indeed, this is the preferred interpretation for English native speakers, as shown by a simple task where participants are asked who had the accident? Spanish sentence grammar allows the same structure, as shown in (10.57). But for Spanish native speakers the preferred interpretation is that the relative clause modifies the first noun la hija (the daughter) (Cuetos Mitchell, 1988).
This difference between English and Spanish is reflected also in reading time studies with sentence materials similar to (10.56) and (10.57), but in which some aspect of the interpretation of the relative clause forces its attachment to one noun rather than to the other. That is, relative clauses were constructed with meanings that more obviously modified either the first noun (daughter) or the second one colonel. If the relative clause modified the first noun then the English participants experienced a gar den path, if it modified the second noun then the Spanish participants experienced a garden path.
Subsequent studies have extended the investigation of this ambiguity type to a range of languages that allow the same sort of structure. It turns out for example that Swedish, Norwegian and Romanian 9Ehrlich, Fernandez, Fodor, Stenshoel & Vinereanu, 1999) as well as Arabic (Abdelghany & Fodor, 1999) follow the same pattern as English, with a preference for the relative clause to modify the second noun phrase, while native speakers of French (Mitchell, Cuetos & Zagar, 1990 ; Zagar, Pynte & Rativeau, 1997), German (Hemforth, Konieczny & Scheepers 2000a, 2000b ), Dutch (Brysbaert & Mitchell, 1996) and Greek (Papadopoulou Clahsen, 2003) side with Spanish, and prefer to interpret the relative clause as modifying the first noun phrase. In addition, however, it has been found that structures where there are three noun phrases that are candidates for modification by the relative clause, such as (10.580 for English and the translation equivalent (10.59) for Spanish, lead to the same results for both languages, namely a preference for attachment to the most recent noun (house/casa).
To account for these different results for sentences with two vs three possible antecedents in the two languages, it has been suggested that there are two competing parsing strategies, and that the cost associated with one of these differs across languages (Gibson, Pearlmutter, Canseco Gonzalez & Hickock, 1996). First, there is a Recency Principle similar to Late Closure, and which is found across all languages. This simply states that there is a preference to attach material to a recent point in the pre ceding structure or alternatively that there is an increasing processing cost with more distant attachments. The second principle, Predicate Proximity, results in a preference to attach the relative clause to the noun phrase that is closest to the verb in the main clause, or that is otherwise highest in the structure (i.e. to daughter/hija as the object the verb in (10.56) and (10.57), and to painting/pintura in (10.58) and (10.59). The costs associated with violating this principle vary across languages, and this has to be learned through exposure during acquisition. In some languages – such as Spanish – the cost of violating Predicate Proximity is high enough to result in it being preferred over the Recency Principle in the two-antecedent cases like (10.57). However, when the highest antecedent is more distant from the relative clause, as in the three-antecedent case in (10.59), then the cost of violating Recency increases, and is greater than that of violating Predicate Proximity, and so the most recent ante cedent is preferred. Additional research on second language learners suggests that – regardless of their first language and of the target language involved – these participants are more likely to make parsing decisions based on Recency. This suggests that Recency is indeed a more fundamental principle.
الاكثر قراءة في Linguistics fields
اخر الاخبار
اخبار العتبة العباسية المقدسة
