Barriers to comparison of socially meaningful objects (groups, behaviors, attitudes,
etc.) across different contexts are an ongoing concern, as attested by advances in
research aimed at detecting and overcoming them. This paper presents problems stemming
from structural variance (incongruity) of discourse contexts which, in addition to
pertaining to different socio-cultural groups are also separated by a century plus
gap: Press coverage of, respectively, late-19th Jewish and early-21st century Muslim
migrants. Their study involving geometric data analysis displays the crux of the apples-to-oranges
problem in that the investigation seeks to find (a) in a joint space of the two groups
(b) discourse clusters (frames) capturing distinct logics of representation and (c)
the latent dimensions in which those frames get articulated. A commonsensical yet
mistaken route would be to isolate those clusters in a "pancultural" analysis ― that
is, in a joint sample but ignoring the two subsets ―, as some of the discourse frames
isolated without considering the groups represented (Jews versus Muslims) might not
be found in their respective subsamples. As a result, the researcher might gloss over
frames that only exist in discourse on one of the groups. On the other hand, when
looking for group-specific frames within separate samples, another issue arises in
terms of dimensionality since the latent forces structuring the discourses in the
two subsamples are likely to be incomparable (if anything, at the metric level). Nonetheless,
innovative methods from the geometric data analytical toolkit make it possible to
both reveal group-specific discourse clusters and locate these along dimensions that
are common to the groups being compared. The implications include mechanisms of racialization
specific to Jews v. Muslims: Like attitudes in general, racialization is relational,
hence impossible to apprehend with reference to a general template.
