In DGS, attitude role shift is typically used to express utterances or thoughts of other people, that is, the signer reports linguistic actions. Therefore, the material used in attitude role shift is mainly linguistic but it can also involve paralinguistic aspects such as facial expressions and signing style. This is illustrated by the following two examples. In example (a), the signer reports that in another discourse context, a signer, namely Tim, asked an addressee, namely Anna, whether she is sad. This kind of role shift equals direct speech in spoken languages in that the reported material is repeated almost verbatim (including the relevant grammatical non-manuals marking a polar interrogative [Syntax 1.2.1]). The same holds true for the second example in (b). Here, the signer reports the assertive speech act [Pragmatics 3] that Tim will help Anna tomorrow. The corresponding declarative clause [Syntax 1.1] tomorrow 1help2 is not marked non-manually.
rs:3a
y/n
a. t-i-m ix3a a-n-n-a ix3b ask ix2 sad ix2
‘Tim asked Anna whether she is sad.’
rs:3a
b. t-i-m ix3a a-n-n-a ix3b tell tomorrow 1help2
‘Tim said to Anna that he will help her tomorrow.’
(Herrmann & Steinbach 2012: 211)
In both examples, attitude role shift is accompanied by non-manuals, which are aligned to the loci in the signing space the reported signer (i.e. ‘3a’) and the reported addressee (i.e. ‘3b’) have been linked to (i.e. the referential loci). The non-manuals typically accompany the reported utterance but may already start in the matrix clause in clause-final position at the speech act verb. In DGS, the following non-manuals are used to mark attitude role shift overtly:
i) Eye gaze change towards the R-locus the addressee of the quoted utterance has been linked to in the previous discourse.
ii) Change of head position towards the R-locus the addressee of the quoted utterance has been linked to in the previous discourse.
iii) A body shift towards the R-locus the addressee of the quoted utterance has been linked to in the previous discourse.
iv) Body lean including a sideward movement of the upper part of the body towards the R-locus the signer of the quoted utterance has been linked to in the previous discourse.
v) Facial expression associated with the signer of the quoted utterance. The facial expression is a gestural imitation of the specific features of the quoted signer relevant for the current discourse.
While the first three non-manuals (eye gaze, head position, and midsagittal body shift) are oriented towards the R-locus of the reported addressee, the fourth non-manual (body lean) is aligned with the R-locus of the reported signer. In contrast to these four non-manuals that depend on the R-loci assigned to signer and addresses in discourse (i.e. on grammatical features), the last non-manual does not depend on grammatical features but on specific extralinguistic properties of the quoted signer.
Taken together, attitude role shift is realized simultaneously by more than one non-manual marker. However, the multiple realization of attitude role shift is not obligatory in DGS. Especially body lean and head movement are less frequently used than eye gaze and facial expressions. The following example illustrates the simultaneous use of all five non-manual components in attitude role shift. Since the signer reports a small conversation between little Emma and her mother, the body lean does also involve a slight upward (Emma) or downward (mother) movement.
rs:3b
fe, eg, bl
hs hs
e-m-m-a ix3b mother ix3a e-m-m-a ix3b ix1 tell ix1 ixa stay play wish
‘Emma said to her mother: “Hey, I don’t want to stay here and play.”
(based on Herrmann & Steinbach, 2012: 215)
However, in many examples, not all four non-manual markers are used. A minimal realization of attitude role shift may consist of eye gaze or facial expression only. This is illustrated by the following example, in which the signer reports the same utterance of Emma without using body lean and head movement, i.e. in this example, attitude role shift is marked only by eye gaze towards the addressee.
rs:3a
eg
e-m-m-a mother 3atell3b ix1 long play ixa wish.not
‘Emma goes to her mother says: “Mum, I don’t want to play here any longer.” ’
Eye gaze and facial expression are the most frequent non-manual marker of attitude role shift in DGS. Body lean and head movement are less frequent and typically combine with eye gaze and facial expressions. This means that the signer uses one minimal non-manual marker to indicate the reported addressee (i.e. eye gaze) and one non-manual marker to indicate the reported signer (i.e. facial expression).
The previous example illustrates that attitude role shift can be introduced by a main clause (i.e. “Emma said to her mother …”) containing signer and addressee and the respective speech act verb (i.e. ask, say, or tell). However, matrix clauses are not necessary to introduce role shift. Moreover, the matrix clause can only consist of the speaker and addressee but need not contain a speech act verb. This is illustrated by the example below showing the dialog between Emma and her mother. In this example, the matrix clause only contains the signs emma, mother, ix3b, which are used to introduce the two interlocutors of the reported conversation. The relevant speech act verb can be omitted because the corresponding speech act is expressed by the reported utterance.
rs:3b
fe, eg, hp, bl
hs hs
e-m-m-a ix3b mother ix3a e-m-ma-a ix3b ix1 tell ix1 ixa stay play wish
‘Emma said to her mother: “Hey, I don’t want to stay here and play.”
(based on Herrmann & Steinbach, 2012: 215)
Another important aspect of attitude role shift is the interpretation of indexical expressions (i.e. ix, here, and tomorrow). In DGS, first and second person indexicals (i.e. ix1 and ix2 and the corresponding inflected forms of agreement verbs) are obligatorily interpreted in the scope of role shift, that is, they are resolved in the context of the reported or quoted utterance. This is illustrated by the examples above. By contrast, temporal and local indexicals tomorrow and here are more flexible in role shift and can be interpreted in both contexts.