Martin (1992: 108):
LEAST CENTRAL #||$ He asked any guy he met there. (pl. — any guys) He asked /some/ guy he met there. (pl. — guys) He asked one guy he met there. (pl. — /some/ guys) He asked this guy he met there. (pl. — these guys) MOST CENTRAL
The outer poles of this scale involve the features [unrestricted] and [major role]. The option [unrestricted] is realised by any, as it is with specific presenting pronominal groups (i.e. anyone, anybody, anything), and signals that it does not matter which members of the experiential class in question is being referred to. At the other end of the scale, the feature [major role] is realised by the demonstratives this/these in more "spoken" registers and a certain/certain in more "written" ones.
 Having previously claimed (p107):
The more central the participant the more likely it is to be Theme, the more likely to be Agent or Medium …
Martin here presents examples of different degrees of centrality, in which the nominal groups in question all serve the same textual (Rheme) and experiential (Beneficiary) functions, and in which the more central participant is neither Theme nor Agent nor Medium.
 As previously noted, a scale that features [unrestricted] and [major role] is invalid because it confuses a deixis category [unrestricted] — interpersonal metafunction — with the "centrality" [major rôle] of an instantial participant — experiential metafunction — in a text.
 This continues the misconstrual of interpersonal deixis as textual reference, and the misconstrual of non-specific deixis (any) as unrestricted reference in specific nominal groups.
 This continues the confusion of presenting participants with referring to referents.
 This confuses reference to referents with the deictic classification of Thing; Halliday & Matthiessen (2014: 365):
The Deictic element indicates whether or not some specific subset of the Thing is intended; and if so, which.
 To be clear, this last example is the only one in which a Deictic also functions as a reference item. Non-specific Deictics (any, some, one), by definition, do not specify an element as identifiable.
 To be clear, the post-Deictic certain does not function as a demonstrative reference item; see Halliday & Matthiessen (2014: 626).
 One the one hand, the claim that one form (determiner) is used in spoken MODE and the other (adjective) in written MODE is a bare assertion, unsupported by data of argument. On the other hand, the claim is demonstrably false, since the two forms do not serve the same function; see  and .