Do speakers treat Mandarin and Cantonese as languages or dialects?
Speaker: Zhenguang Cai
Abstract:
Meaning can be expressed in different codes (e.g., English or Dutch). The distinction of two codes as two languages or two dialects of the same language is often conventionally rather than empirically defined. We used structural priming to empirically investigate this issue. Research shows that there is a boost of priming effect (lexical boost) when the verb is repeated1 and a translation-equivalence boost when translation equivalent verbs are used in cross-language priming, though weaker than the lexical boost due to the fact that translation equivalents are represented as different lemmas2. This provides a good way to test whether equivalent verbs in two related language varieties share the same lemmas (and thus whether they belong to the same language): if they do, there should be equal boost in within-language priming and in cross-language priming when the verb is repeated; and if they don’t, cross-language priming would result in translation-equivalence boost only.
We conducted an experiment (128 participants and 32 target items) to investigate whether Mandarin and Cantonese (which share most of the grammar but are mutually unintelligible) are the same language. Participants decided whether a prime sentence involving a double-object (DO) or prepositional-object (PO) structure matched a picture and then described an unrelated target picture that could be described using a PO or a DO structure. We manipulated 4 factors: Target Code (half of the participants described the picture in Mandarin and the other half in Cantonese), Verb Meaning (whether the prime and the target have same-meaning or different-meaning verbs), Code Consistency (whether the prime and the target involve the same code or different codes), and Prime Construction (whether the prime has a DO or PO structure). Results show that 1) there are more primed responses (DO response after a DO prime and PO response after a PO prime) than unprimed responses; 2) Mandarin as a target code induced more priming than Cantonese; 3) same-meaning verbs induced more priming than different-meaning verbs; 4) same-code priming induced more priming than different-code priming; and 5) most interestingly, there was an interaction of Verb Meaning and Code Consistency. The interaction shows that the boost with same-meaning verbs was stronger in same-code priming (Mandarin to Mandarin or Cantonese to Cantonese) than in different-code priming (Mandarin to Cantonese or Cantonese to Mandarin), indicating that the boost in the same-code priming was a lexical boost while the boost in the different-code priming was a translation-equivalence boost.
These results support the view that Mandarin and Cantonese are represented as two different languages rather than two dialects of the same languages. We also discuss the implications of the present study for cross-language structural priming.
References
1 Pickering & Branigan (1998). JML, 39, 633-651.
2 Schoonbaert, Hartsuiker, & Pickering. (2007). JML, 56, 153-171.
