138 andrew w. pitts
these features, gives each new testament book what he refers to as an
“involved-informational Production” score. letters can receive a score
that is more or less (±) interactive, or more or less (±) informative. the
± scale allows for precision and can model for readers even subtle varia-
tion within register profiles at the level of interactiveness.
the Pauline corpus groups into essentially 4–5 register profiles (depend-
ing on the origin of galatians) based on several significant time slices in
Paul’s timeline that provide the significant social settings for letter compo-
sition. We immediately notice two very different sets of variation between
two groups of register profiles. register profiles 1–4 (the ten ecclesial
letters) group together in several ways that make them distinct from reg-
ister profile 5 (the personal or Pastoral letters), thus exhibiting the more
significant variation. there is substantial register variation within the
mode of discourse. in the primary channel component, we observe less
formalized epistolary structure in profile 5. our landscape also picks up
a substantial shift in format/permanency. Profiles 1–4 are at least semi-
published, that is, published locally within a church or network of house
churches or intended to be circulated throughout a region. By contrast,
profile 5 was apparently originally intended to be unpublished. the set-
tings also differ significantly with profiles 1–4 having a semi-public, but
still personal, locally circulated setting whereas profile 5 shifts strongly
toward privacy and, therefore, intimacy, as well as a likely shift toward
specialization due to the teacher-disciple relation. i explore further varia-
tions below in connection with Pauline language variation.
A Register Design Interpretation of Pauline Language Variation
Based on this register profile landscape, we have a scale of social cluster-
ing and variation that can track social change in tenor alone.
this graph correlates the level of co-textual variation with contextual
variation in order to assess their correspondence, as our register design
model of style predicts.69 the graph only models the interpersonal fea-
tures, since these tend to have the greatest impact upon style-shift and they
help to group the register profiles with clarity. register profiles 2–4 group
tightly with addressee type but vary somewhat on involved-information
69 Since southern and northern galatia theories cannot be true, i have chosen to include
galatians with register profile 3 (rather than 2), since that corresponds more closely with
scholarly consensus. Profile 2 includes only the thessalonian correspondence.