Current research into second-language learning has tended to ignore (or at best to treat incidentally) a linguistic phenomenon that once used to be a particular preoccupation of applied linguists, the interference error. Instead, the limelight is now firmly focussed on developmental phenomena, with many studies using an approach to data gathering and analytical methodology strongly reminiscent of research into child language acquisition and language contact. There have been specific attempts to establish developmental sequences in the TL, in morpheme acquisition, for instance, so as to compare first and second language learning, and not a little attention is now being paid to such sociolinguistic notions as variability, continua and simplification. In other words, the main emphasis in interlanguage research has shifted from a rather static error-oriented view of language learning to a dynamic view of learners' language as a constantly evolving system. The calls for longitudinal studies of interlanguage of the late sixties and early seventies have not gone unheeded, eve if the word ‘longitudinal’ is sometimes rather liberally interpreted, as in those cases where a tacit (and probably justifiable) assumption is made that studying groups of learners of varying proficiency in respect of given TL features at one and the same time is really the same as following the progress of one group over a long period.