Cross-Linguistic Suffix Preference: Typological or Cognitive Bias?

The origin of a typological suffixing bias-the tendency to exploit suffixes more often than prefixes to express grammatical meanings-is unknown. One hypothesis suggests that there is a cognitive bias favoring sequences with a variable part appended at the end of a fixed part rather than sequences with a variable part appended before a fixed part. We aimed to determine whether such cognitive bias exists and whether it is domain-general or language-specific. We set up an artificial language learning experiment with native speakers of Arabic. In Arabic, grammatical meanings are expressed via vowel templates rather than by adding a variable ending (suffix) or beginning (prefix) to fixed syllabic constituents (stems). This allowed us to minimize the effect of native language typology (whether it uses suffixes or prefixes) on cognitive performance. We did not observe evidence of a cognitive bias in Arabic speakers and suggest that the individual cognitive bias observed in earlier studies is defined by the typological properties of participants' native languages. This casts doubt on the possibility that domain-general cognitive bias could lead to the emergence of typological suffixing bias across world languages. We propose several possible origins of typological bias; each requires empirical verification.