The recent paper by Spencer and Dixon (1983) reports the successful application of Wood's experimental mapping language, Environmental A, for capturing feelings and impressions about the urban environment. Several cautionary comments may be made on the authors' advocacy of wider use of the technique' in perception studies so that the empirical tradition may begin to 'incorporate aspects of the humanistic-phenomenological approach' (p. 382). First, preliminary discussion on the neglect of the affective dimension in cognitive mapping exercises hitherto would benefit from some clarification. Here, as often in the literature in general, there is confusion between perception and behavioural studies, whether one is a subset of the other. Thus, while 'the argument' presented is 'that the study of ... essentially spatial and locational aspects of the cognitive map is not sufficient to predict ... individual choices' (p. 347), their paper is solely concerned with mapping. Moreover, it is not strictly correct to say that conventional cognitive mapping exercises 'lack any evaluative (or affective) dimension' (p. 347). Objects are not recalled for their physical presence alone but because they are significant for something-functionally, socially or symbolically, the particular reason being determined by the instructional set. Thus, for instance, Milgram (1977) mapped the affective response of Parisians to their city from the itinerary and landmarks chosen for a final walk in their capital before being sent into imagined exile. Again, the whole group of cognitive maps concerned with the evaluation of places in terms of residential desirability, perceived beauty, stress, danger, pollution or whatever, illustrates the integrated nature of the response. It is unfortunate that the discussion of humanistic approaches (sense of place, rootedness, insideness are listed) concludes with the claim that 'such subjective elements can be incorporated into the main empirical tradition' (p. 375). The 195 symbols of Environmental A may well be superior to an adjectival check-list, but the methodology remains one that constrains the respondent, quite unlike that in humanistic studies. Quite simply, different philosophies are involved. Here, however ingenious the symbols, responses are selected from a pre-packaged set, they are not spontaneously generated. Even when involvement and habituation have replaced anticipation and encounter with a particular environment, common or universal experiences may remain untouched-strength and consistency of feeling, even articulation or expression itself. The experiential dimension of humanistic work-for which the term 'feelings' is perhaps more appropriately applicable-is much richer than the response presented here, which is more one of evaluation.