St Bridget of Sweden (1303-73) had a remarkable and eventful life, summarized by Bergh (1981) but not mentioned by Karlsen in the study under review, during which she had visions of Christ and the Virgin Mary. She eventually recounted some of them in writing in Swedish, and some of her Swedish colleagues, most notably the theologian Mathias of Linköping, translated them into Latin. She moved to Rome in 1349, where the texts were revised for public consumption, perhaps twice, by her confessor Alfonso, bishop of Jaén in Northern Andalucia (regularly mis-spelt in this book as "Alphonso"). It is thus not at all clear who the individual linguistic features of the text should be ascribed to. Bridget herself may not have been capable of composing much in Latin; the Latin written by her compatriots was a foreign language, probably moulded for the most part by their grammatical instruction; the Latin of the final version, polished to an unknown extent, was that of a Romance-speaker whose native linguistic instincts necessarily underlay the high style he adopted for professional purposes. So there could hardly be a text less clearly representative of the idiolect of its original author (and if it represents an idiolect at all, it is probably that of Alfonso of Jaén). Nonetheless, Karlsen ignores all these problems and cheerfully bills his subject matter in this book as "a limited and homogeneous text corpus" (138), and the linguistic data he unearths as ". . . phenomena that must be regarded as fundamental rules of Bridgettine grammar" (142). What an odd comment; "Bridgettine grammar", if the phrase makes sense at all, can only have been a kind of Swedish.
In the event, we don't find "fundamental rules" here either, but statistics. Statistics can, of course, be true, and Karlsen has worked his out with such careful diligence that his readers will be happy to believe that these are. This does not necessarily mean that they have more ontological validity than the phenomena they thereby summarize, so the fact that in these Revelations indirect statements are presented as quod- or quia-clauses on 55% of the relevant occasions and on the other 45% as non-finite constructions, predominantly accusative and infinitive, indicates nothing very much in itself; but Karlsen has come up with statistical correlates that may indeed explain authorial choices between these two alternatives. The main correlation presented is that the more "assertive" the main verb is, the more likely the subordinate clause is to be introduced by the complementizer quod or quia (or even quasi, in the case of parables). "Assertion" is a most interesting concept whose relevance to grammatical analysis is currently much discussed by theoretical linguists; Karlsen does not wish to enter this terrain, however. Indeed, if the readers even wish to know how this and other technical terms can be defined, and what is involved in their use, Karlsen has decided not to enlighten us himself but to refer us instead to a fairly obscure 1994 Italian dissertation by Pierluigi Cuzzolin (43 n. 30, 62 n. 54)). There must be several hundred more enlightening and accessible sources for readers of a book in English than this. Furthermore, he lets on that the subject of the "assertive" main verb is often Christ or Mary, so maybe "authoritative" would be a better description, and in the event more illuminating, in that the indicative in the subordinate clause after quod tends to convey more of a factual implication than the necessarily mood-neutral infinitive (even though Karlsen points out (80) that quod plus subjunctive, which is quite common, does not necessarily imply a lesser view of the relative truth of the complement on the part of the narrator).
Verbs of perception are included in this analysis. They are more likely to be used with the accusative and infinitive than with quod. Again, whether we think that the careful statistics (23-24) of 69% ( quod with indirect statements) versus 45.5% ( quod with verbs of perception) are significant, rather than just being a restatement of the data, is debatable. In the event, if Karlsen had been using his data for some interesting purpose, such as determining the relative authorial input of Bridget, Mathias and Alfonso, these differences between perceptive and declarative complements might incline us to attribute a significant role to Alfonso; for in the Ibero-Romance that underlay his written language, indirect statements normally had a complement introduced by que, but verbs of perception regularly had, and indeed still have, the accusative and infinitive (or, more often in Modern Spanish, the infinitive and accusative: e.g. vi comer a la madre, "I saw the mother eat", where the a marks the mother as being the object of the seeing while being at the same time the subject of the eating). Unfortunately, Karlsen is sternly uninterested in such sociophilological concerns. Indeed, he even deliberately avoids analysing in detail the rubrics supplied to the text by Alfonso, which could form a valuable control group. It seems that all possibly interesting conclusions leading from the statistics are intentionally left unexplored. The possible influence of the Vulgate on the Revelations might seem ostensibly an obvious and important area to consider, but it is described here as being merely "difficult to ascertain" rather than actually being investigated (27). The accusative and infinitive construction existed in the Swedish of the time, after both categories of verb, but the possible influence of these constructions on the surviving text is merely described as being "difficult to decide" rather than studied seriously (28). They could be crucial, of course, in indicating whether Bridget had more linguistic input into the final text than otherwise seems likely.
The statistical truths that are discovered in this analysis concern for the most part an increasing preference for quod-clauses correlative to greater syntactic complexity in the sentence as a whole. This tendency could be compared by an analysis with the pedagogical grammars probably used for the training of the Swedish intermediaries, for if it corresponds to reconstructable technical advice mentioned in their professional formation, then we could assign an important role in the text to Bridget's Scandinavian colleagues; if not, this would be a further argument for a fundamental role played by Alfonso in the text that survives. The fact, discovered by Jozsef Herman, that in Late Latin quod-clauses always follow the main verb while the accusative and infinitive construction has no relative positional preference, further supports the latter's role, since the Revelations follow the same pattern and Herman, as Karlsen refrains from pointing out, was analysing texts written by native speakers (whose speech would have post-verbal quod). It also seems likely that there was often a pragmatic basis for the choice, in that quod-clauses were more often used for new information and the accusative and infinitive for information that the listener may well have already known. If Karlsen had been interested in pragmatics, or discourse analysis, or the use of language in context, he could have used this tendency to bolster his assertion that assertion is the key criterion for the choice of quod, since new information usually needs assertive foregrounding and old information usually does not.
This review is not meant to sound negative. The statistics have been worked out thoroughly and presented professionally according to a wide range of parameters, including the lexical verb (some verbs definitely prefer one construction to the other) and the syntactic circumstances of the main verb. There are two long chapters on comparatively detailed sub-topics; the rendering of the future as seen from the past, i.e. "would" (using the "prospective imperfect subjunctive", 94, chapter 7), and the constructions that are used after verbs of wishing and allowing where the infinitive might or might not be merely prolative after a nominal object (chapter 9): is admoneo te venire "I warn + accusative and infinitive" or "I warn you + infinitive"? (126; the answer is essentially that already provided by Pinkster).
Philologically this analysis is admirable, in the best Scandinavian tradition. But it is probably time now for the philologists to analyse texts in their sociocultural context as well as internally (which is why in my view there is a need to establish a new discipline of sociophilology). This is not just because sociolinguistics is fashionable; the connection can be illuminating in both directions. An understanding of the precise circumstances of composition can help us, among other things, explain some of the philological details discovered by valid internal analyses such as this. (This is not the "Intentional Fallacy"; historical linguistics and philological analysis are not easy, and it is no sense cheating to use all the contextual evidence that we can.) Conversely there are texts where the internal analyses can actually clarify the circumstances of composition, and this can still be the case with St Bridget's Revelations; Karlsen may well now come up with interesting further conclusions on the basis of the data here presented, but so far this book has a preliminary air. It is a doctoral thesis, apparently presented neat, but it really deserved expansion before publication.
References:
Bergh, Birger, 1981. "A Saint in the Making: St Bridget's Life in Sweden (1303-1349)", Papers of the Liverpool Latin Seminar, 3, 371-84.
Cuzzolin, Pierluigi, 1994. Sull'origine della costruzione DICERE QUOD: aspetti sintattici e semantici, Firenze: La Nuova Italia.
Herman, Jozsef, 1989. "Accusativus cum infinitivo et subordonne' quod, quia en latin tardif - nouvelles remarques sur un vieux proble`me", in G. Calboli, ed., Subordination and Other Topics in Latin, Amsterdam: John Benjamins, 133-52.
Pinkster, Harm, 1990. Latin Syntax and Semantics, London: Routledge.