Hammersley (200*) criticises a particular style of discourse research for developing as a distinct paradigm, yet lacking the coherence a paradigm would require. He suggests a range of problems in relation to constructionism, reflexivity, and the 'thin' model of the human actor, and argues instead for methodological eclecticism where discourse analytic methods are supplementary to alternatives. This commentary highlights a range of confusions and misunderstandings in this critique. In particular, it highlights the way discourse analytic work is connected to a range of theoretical notions, most fundamentally in its theorising of discourse itself as a medium oriented to action. It identifies important sources of incoherence that can arise when mixing discourse analytic and more traditional methods. It reiterates the virtues of constructionism, particularly when considering the operation of descriptions, stresses the value of exploring (rather than ignoring) reflexive issues, and emphasises the rich and nuanced approach to psychology that has been developed in this tradition.