One of the most important issues in the study of language processing concerns the universality of sentence processing strategies across different languages (e.g. Hillert, 1998; de Vincenzi, 2000). As all researchers agree, if the goal of psycholinguistics is to study the human sentence processor and not the sentence processing mechanisms of a specific language, the test of universality is not an option but a necessary condition for evaluating various models of sentence processing.
Crosslinguistic comparisons of sentence processing can be attempted from two different perspectives. One is the universality hypothesis. It argues that there are universal processing strategies that apply to all languages, because processing strategies are independent of specific languages and are based on cognitive universals. The strategies or principles that could be considered as being universal are minimal attachment and late closure (Frazier, 1987;) as well as the minimal chain principle (de Vincenzi, 1991). These principles are assumed to operate in all languages, with only the ‘vocabulary,’ i.e. lexical items and specific grammar, differing across languages.
Another reason to do crosslinguistic studies might be the expectation that different languages show different processing strategies, under an assumption that parsing strategies are a reflection of language-specific characteristics and a by-product of exposure to a given language. Some processing strategies may not be universal but are instead language specific or parameterized (e.g. Mazuka, 1998). So, there is no need, at least in principle, to test the same strategy in different languages.