The paper recaptures, on the basis of one of the central issues of the discussion, namely, the relationship between thought and speech, the psychlin-guistic controversy between Wilhelm Wundt (1832–1920), Hermann Paul (1846–1921), and Anton Marty (1847–1914) at the turn of this century. The basic tenets of all three theories are presented, their assumptions analysed, and their respective fruitfulness (or lack of it) put forward. After redressing the distorted picture of Wundt’s position in the recent historiography of psycholingu-istics, it is shown that Wundt’s model of an expression-oriented approach, which in effect identifies categories of linguistic surface structure with those of an inner psychological nature, remains circular and not amenable to further development. Hermann Paul, though making use of a similar procedure, is opposed to Wundt’s (as well as Heymann Steinthal’s (1823–1899) social psychology or Völkerpsychologie), favouring instead the individual as the locus of linguistic events (and hence linguistic analysis), thereby playing down the importance of linguistic intercourse and communication in language acquisition and historical development. Finally, in Marty’s theories the contradiction between his reliance on 19th-century event-directed psychology and a rather modern functional conception of language is most evident. Marty wants, unlike Wundt and Paul, to distinguish clearly between genetic and systematic questions. But while recognizing the complementarity of event expression and control of comprehension on the part of the hearer, he does not do so in the case of the linguistic representation of ‘objects and events’. In an attempt to escape from the naive homology of thinking and grammar, Marty argues in favour of a complete separation of the two mental activities. The paper argues that the common psychological premisses of these authors must be considered if the differences between them are to be understood, since it is just these particular premises that lie in the way of an adequate comprehension of problems of semantics and of communication.