World knowledge can rapidly inform language comprehension (Ferretti et al., 2001; Pearlmutter et al., 1999) and so can social-context information (Nieuwland & Van Berkum, 2006). Other context-related aspects, like formality, may also matter for language variability (e.g., Adger, 2006). We hypothesize that formality can guide the processing time of ‘variants’ of words and grammar (increased for mismatches over matches). For instance, depart might be expected to reference the act of leaving in formal contexts; in less formal exchanges, we might expect to read scoot or scarper (BE). We investigate whether: (i) context formality- register congruence elicits rapid effects; (ii) variability in formality ratings modulates such effects; and (iii) sensitivity to formality-register congruence and to grammar knowledge interact during comprehension. Point (iii) explores additivity versus interactivity of the implicated mechanism and/or representations.

Methods: In two eye-tracking pilot studies, monolingual German adults read target sentences preceded by two context sentences. 40 critical items were interleaved with 56 filler items (of which 75% were followed by binary comprehension questions). In a register-only pilot (N=8) we manipulated register congruence of context and target sentence (match vs. mismatch, Table 1). In a register-by-morphosyntax pilot (N=8), we additionally manipulated subject- verb morphosyntactic congruence (match vs. mismatch). Linear mixed-effects models were fitted to the log-transformed (exponentially back-transformed for visualization) first-pass duration, regression path duration, and total time at the verb and object (NP2) regions in the target sentence (Table 1). Context and target sentence formality was normed via independent Web-based ratings.

Results: In the register-only pilot, at the verb region, longer total times (Figure 1) emerged for register mismatches (vs. matches, t=-2.05, p=.040, d=-.25), and for higher target sentence formality ratings (t=3.22, p=.001, d=.37; post-verbal effects n.s.). In the register-by- morphosyntax pilot, we expected that the effect of subject-verb morphosyntactic congruence effect might modulate that of register congruence or that the effects might be additive, reflecting, respectively, shared versus two distinct mechanisms. At the verb region, as expected, total times (Figure 2) were longer for morphosyntactically incongruent (vs. congruent) verbs (t=-1.98, p=.048, d=-.24). For register congruence, contrary to predictions, verb total times were longer for matches (vs. mismatches, t=3.24, p=.001, d=.39; effects of formality ratings n.s.). At the NP2 (spillover) region (Figure 3), the analyses revealed effects of subject-verb morphosyntactic congruence (t=2.07, p=.038, d=.25), register congruence (t=- 2.50, p=.012, d=-.31), as well as of their interaction (t=2.00, p=.045, d=.25). Nouns following a morphosyntactically congruent verb yielded longer first-pass durations in the register match condition, whereas in the register mismatch condition no difference was observed between morphosyntactic matches and mismatches. Finally, longer NP2 regression path durations (Figure 4) were observed following morphosyntactically incongruent verbs (t=-2.83, p=.005, d=-.35), register-congruent verbs (t=2.55, p=.011, d=.31), and as a function of higher average context formality ratings (t=2.45, p=.014, d=.29).

Preliminary results of both studies suggest that social-context information (register congruence and perceived formality) and grammar knowledge are incrementally integrated during sentence processing. Interestingly, both effects appeared to emerge at a later processing stage in the verb, but at earlier stages in the spillover region (register-by- morphosyntax pilot only). When both factors were manipulated, the direction of the register congruence effect was opposite to that predicted, even when an interaction was not present. Pending replication with a larger N, these findings suggest some degree of interference between the effects of register and morphosyntactic congruence, potentially relying on not entirely distinct cognitive mechanisms.