Martin (1992: 553):
Beyond the problems of sequence and order reviewed above, the problem of non-discrete realisations which smear across rather than mapping onto elements of schematic structure needs attention. These have been as much ignored in grammar as in genre … and so once again it is linguists' reliance on old tools for mapping new territory that stands in the way of progress …
 The "problems" identified by Martin derive from a number of misunderstandings; see previous post here.
 This is a false opposition. As Halliday & Matthiessen (2014: 166-7) point out, the prototypical example of the prosodic mode of expression that is characteristic of the interpersonal metafunction is tone. However, tone choices do "map onto" the elements of structure of the tone group: the Pretonic and Tonic segments. In doing so, they also "smear across" the feet that realise those segments.
 The implication here is that relying on new tools, such as those that Martin might provide, will not "stand in the way of progress". See the rest of this blog.