摘要：Can sound changes be predicted? And do they optimize systems, as is often assumed? In
this paper we investigate these questions by looking at a still ongoing sound change in
Austrian German: the monophthongization of the diphthongs /aε,/ and ,/αο/. Based on a
detailed acoustic phonetic description of the changes of the diphthongs since the late fifties
we show that sound change is triggered by principally predictable circumstances, but its
outcome cannot be predicted due to the complexity of the competing (intra- and extralinguistic)
factors involved. It is argued that models of evolutionary adaptation cannot
account for these facts, whereas the theory of 'autopoietic systems' (Maturana & Varela
1987) is able to explain the seemingly unpredictable behaviour of the language system.
