California Linguistic Notes Volume XXXV No. 1 Winter, 2010
California Linguistic Notes Volume XXXV No. 1 Winter, 2010
California Linguistic Notes Volume XXXV No. 1 Winter, 2010
This work constitutes a veritable mountain of scholarship, consulting primary and secondary
research from the early 1800s to the present. Chapters take up in turn inscriptions from the
republican period, explicit evidence for regional variation in the early Republic, evidence from
the Empire, region by region, i.e., Rome and Italy, Spain, Gaul, and Africa, provincial texts from
Gaul, from Spain, from Italy, from Africa, and from Britain, and finally inscriptional evidence
across regions spanning the empirical period. Abundant examples are cited, and the Latin data
are translated to English. As the data are developed and accounted for, significant contributions
are made not only in the immediate topic, but in determining sound methodologies for such
investigations. The text is supplemented in the forward matter with an index of maps and a table
of primary source references, and in the after matter with 18 maps, an index of words, and an
index of references to primary sources, along with the usual bibliography, reference pages, and
subject index.
The author is absolutely correct that a neat history of Latin and its development into
Romance cannot be written (xvi). And that finding evidence of regional variation outside the
historical work in Romance and later Latin is dominated by a comparative perspective, i.e.,
looking backward. The problem with the view through the rearview mirror, as the label affixed
clearly states, is that objects may look farther away than they actually are. The goal of the
present study is a systematic account, looking forward from antiquity, not backward from
In establishing the perspective for the work, the analogies of English, French, and
German being planted in extranational regions are consulted to hint at processes that might have
occurred during the spread and exportation of Latin, but with the recognition that the parallels
break down at the great difference in time depth (21 27). The question of what constitutes a
dialect is an early, central question in framing the study. This is perilous terrain in this field,
where diachronic variation can appear as regional (9, 693, et al.), given the haphazard character
of the available evidence, the nature of the spread of new features, bilingualism and imperfect
learning of Latin, varying degrees of literacy among scribes and masons, and the great time
depth involved.
of this book is the categories strong and weak regionalisms. Strong regionalisms are those in
which a local term is used while a synonym or near synonym is broadly used outside the region.
Weak regional terms are those that result from the local influence for which the language has no
parallel, such as the names of fauna and flora, mining terms, and the like. These categories may
be exemplified in Australian English, where the diminutive clipping barbie is used where other
Englishes use barbecue (grill), while koala names an animal that lives in the region, for which
no other word in English is available, and which word is indeed in use in every variety of
English for the associated mammal. Frequently such weak regional words are loans from local
indigenous varieties, but this fact is secondary in significance in terms of establishing the
regional character of the variety vis--vis the broader language. This contribution imposes rigor
on the study of variation. Thus, through the accident of geography the language spoken in a
region like Australia can acquire loan words first, which then diffuse through the world English
speaking community, so in hindsight what sets it apart as a variety are those elements which are
One source of evidence in both the Republic and Empire periods involves metalanguage
absence of descriptive observations, of the speech of given regions. These constitute a long-
familiar topos, that of the supposed differences between the speech of urban and rural speakers.
This appears in texts by writers like Varro and Cicero and in satiric theatrical productions in
which rustic speech is parodied and jokes made which pun on pronunciation differences (123, f).
Adams rightly concludes that such variation must have been present, but recognizes the difficulty
in identifying what the specific features that vary must be. Sometimes features are identified, as
when Cicero in some spots reports his impressions of the speech he discusses but also notes that
dropping [s] in certain contexts, which had formerly been a sign of refinement in Rome, in his
Drawing conclusions from such evidence is not without difficulty, as analogous to the
dropping and restoration of [h] in modern English, the social target of change was a moving
one. And even so, evidence is consulted to show that that feature was far from consistent in
available literature (140, 41). Meanwhile, Varro notes specific lexical variants in areas around
Rome in the Late Republic period (173). An fascinating process reported is the evident
rhoticization of [d], which appears as a hypercorrecting tendency in the provinces during the
Evidence for variation from inscriptions is hard won. The orthographic conventions in
relation to phonetic factors must be considered (43), along with evidence for provenance and
authorship. Questions regarding the nature of the text, the degree of literacy represented, and
language affiliation of the author arise. Apparent distinctions in features represented in texts
from Rome and the Italian Peninsula (suffix a or ai in certain grammatical contexts) in fact
reflect nothing more than that the Italian texts consist of material in the religious register, which
consistently differ from secular texts in the feature, even in Rome (46, 47). The appearance of i
or e reflects instability of the vowel in unstressed contexts (244). What is required is patient,
methodical, thorough sifting of the evidence and comparing it to what else is known or being
discovered. Thus the author concludes, for example, that funerary inscriptions can present false
regionalisms in the formulae, common in a locality but not in use in the local spoken language,
that guide their preparation (678). Each of the many items developed in these pages is the
product of considerable refinement, which contributes to our knowledge how to effectively study
the evidence.
this study: its methodology. Each piece of evidence is filtered through what is known of any
other available evidence and what is known about language change, structurally and socially, to
draw conclusions which often are recognized to be quite tenuous. After two generations in
general and theoretical linguistics during which conclusions were informed by the intuitions of
writers, informed perhaps an example or two, here appears a scientifically disciplined method
that permits us to conclude only what we have good evidence for. Throughout the text we are
exemplified by the instability of i and e in unstressed contexts from the earliest stages (244, 45).
This approach yields an apparent isogloss across parts of Gaul and the Iberian peninsula that
marks the distribution of canastrum, a loan, and canistrum, the usual Latin word (283), but
considerable evidence must be consulted to make that determination. The above example also
shows that lexical areas are not identical with political boundaries and language areas (702, et
al.). This painstaking sifting and comparing of the evidence contributes a scientifically rigorous
methodology.
Another contribution is a set of criteria for placing texts of unknown provenance. They
involve words not in use in the literary language, like tructa trout in Galllo-Romance which is
very rare in the literary language of extant Latin (336). Conversely, Adams states, I conclude
that commonplace words with a long history in written Latin can establish nothing about the
place of composition of a text (358). While it is acknowledged that terms belong to localized
mundane words, not items in the literary language, identify a regional speech, as literary Latin
The author likewise argues convincingly that the date of colonization is not the decisive
determinant in dialectical differences among the regions (430, et al.) and that it is a mistake to
look only for dramatic external forces or major historical events (508). After all, good evidence
exists for the mobility of Latin speakers, code mixing, and bilingualism throughout the provinces
(e.g., military units in Spain, 370, and at Vindolanda in Britannia, 581) and in the broad
dissemination of written Latin among the literate classes. It is shown often that archaisms that
were in general use may have haphazardly been retained in one region and faded in others; words
such as these cannot help establish the place a text was produced, but they may provide evidence
archaism, was the main factor contributing to the linguistic diversity of the provinces, which is
especially evident in Africa (565), where, as elsewhere, the language was always in a state of
flux (541). Good evidence for variation is found in changes in suffixes that result in a new form
that becomes a dialect word (509) and in technical terms that do not spread more broadly (453,
512).
Yet another important contribution is a set of guidelines for a statistical method, which
begin with determining what the text(s) represent(s) as a sample (636) and involve quantifying
not only the question features, but the features from which they are supposed to vary (649, f;
summarized 662, 63). This method has identified important mistaken conclusions drawn by
previous researches, and helped establish the aforementioned observation about the instability,
An area of application of the present study that has implications for contemporary
theories of language variation and the development of local varieties has to do with productions
of learners of Latin who had not completely acquired the features they produced. It was observed
that foreign language errors on epitaphs can be attributed to incomplete learning of, for example,
the syntactic system of Latin (618). A snapshot of a particular period of language shift may show
items that color local speech when bilingualism is prevalent, but will disappear when the
substrate language is dead (368). Fossilization and interference may pertain to individual
speakers in a generation, but innovation is continuous (541), and the evidence shows that
interference items in the language spoken by later generations had died out. What is in fact a
transitory diachronic phase, incomplete acquisition associated with bilingualism and language
economic, and social influences involved with major and minor historical events across the arts
are consulted, along with the raw linguistic and inscriptional data, when drawing conclusions.
The monumental study reported in this tome can be characterized as cross-disciplinary, involving
as it does structural linguistics, principles of social and regional variation, and the historical-
comparative tradition, but it would perhaps be more fitting to state that such a remark exposes
the folly of relegating the facets of language study to distinct sub-disciplines. This study of
astonishing compass illuminates a topic which after two millennia had still lacked clarity, while
Robert D. Angus
California State University, Fullerton