Review of Reflexive Marking in the History of French
The book is a study of clause-mate anaphora (i.e. when a pronoun is used for reference to the subject of the clause), and related phenomena, in French. The approach is diachronic, tracking the evolution of usage and the distribution of alternations across time.
Chapter 1 introduces the facts: the difference of distribution of “soi” (‘him(self)’, ‘her(self)’ or ‘them(selves)’) versus “lui”/ “elle” (‘him’/ ‘her’); the use versus absence of “-même” (‘-self’) with those pronouns; and the distribution of intensifiers. The author also acknowledges his tie to the framework of Construction Grammar and proceeds to distinguish between coreference (i.e. a discourse relation) and binding (i.e. a grammatical one: the slots of a construction are coindexed), remarking than a binding construction doesn't preclude coreference. Finally, the importance of reflexives cross-linguistically is stressed (while “brute force reflexivization” (Reuland, 2005), i.e., coindexation marked with ordinary pronouns, could be, and is sometimes, used).
Chapter 2 addresses the theory of anaphora. The author first reiterates the above-mentioned distinction between coreference and binding and then discusses specificity (an important side-issue), noting in particular that a non-specific noun phrase can be reinterpreted as specific as the discourse proceeds. Binding Theory (Chomsky, 1981), associated problems, and tentative solutions (from both generativism and functionalism) are then reviewed. The author's main point is that non-complementarity is the rule and complementarity should be motivated, not the other way around; in other words, reflexive and personal pronouns can normally both be used to denote coreference. It is remarked that non-specific subjects allow binding alone. Finally, the author critiques the position that the “soi”/ “lui” (reflexive/personal pronoun) alternation reflects argumenthood: where a pronoun is allowed to mark coreference to the subject, it signals the slot it fills as an adjunct. But data show, at least in French, that this is a matter of construction: some verbs collocate with reflexives while others do with pronouns.
In Chapter 3, the author analyzes diachronic data, tracking the progressive decrease in the use of “soi” versus “lui”/ “elle” from Old French to the present time. The former is now mostly used for non-specific antecedents; in that respect, “on” (‘one’) (i.e. an exclusively non-specific subject in written French) always triggers “soi”. However, inspection of particular verbal constructions reveals that the use of “soi” is not steadily decreasing, but rather develops on a predicate-by-predicate basis; in some cases, “soi” is still widely used even with a specific subject. Such a finding stresses the importance of constructions when studying the evolution of one item. Finally, the author confirms the rare use of “soi” with plural antecedents, namely due to the semantic properties of the distributive plural, ill-fitted for binding.
Chapter 4 investigates the use of “lui/ “elle” versus the reinforced versions “lui”/ “elle” + “-même”. The author distinguishes reflexive predicates from coindexation of arguments: a predicate can be reflexive without taking a coindexed argument (e.g. ‘John washes’) and the reverse is also true (e.g. ‘John never carries money with him’). Then it is argued that “-même” doesn't mark the pronoun it attaches to, but rather the related predicate, creating a new one with focus on reflexivity. A difference in distribution can be found along the following lines:
(1) Jean pense à lui et moi et ...
John thinks of himself and me and ...
(2) Jean pense à lui-même.
John thinks of himself.
In (1), focus is on the pronoun; John thinking of himself is paralleled by John thinking of other entities. In (2), the entire predicate is focused, and possibly compared with other predicates. That semantic difference is sometimes even more obvious; to say something to oneself is not the same thing as saying something to somebody which happens to be you. However, on the diachronic front, no evolution in the use of simple versus reinforced reflexives can be detected.
In Chapter 5, the author distinguishes (after Siemund, 2000) between three different uses of the intensifiers “lui”/ “elle” + “-même” (when not used as reflexives): adnominal (e.g. ‘The king himself’), adverbal-exclusive (e.g. ‘He wrote the book himself’, meaning he didn't hire a ghost writer), and adverbal-inclusive (e.g. ‘I'm a teacher myself’, meaning ‘I'm a teacher too’). The author argues that these three uses are metonymically related, which is reflected in their gradual apparition over the history of French.
Chapter 6 reviews various approaches to language change, including the author's own hypothesis on rhetorical devaluation (Detges and Waltereit, 2002): speakers use the most “noteworthy” forms, so much so that those forms lose their strength, thus becoming new, unmarked constructions. This hypothesis is then applied to the evolution of negation in French and the rise of compound past tenses in various languages. The author then returns to the evolution of “soi” versus “lui”/ “elle”. The alternation is, at least in Old French, a matter of “Differential Object Marking”: the object of the verb can be marked in different ways (e.g. Spanish unmarked “a”-marked objects), according to semantic and contextual criteria. Here, the difference is that “soi” is less thematic than “lui”/ “elle”, since binding is a grammatical relation, whereas coreference happens at the level of discourse. Accordingly, “lui”/ “elle” is stronger than “soi”, hence more “noteworthy”, and was used more and more often until it became the unmarked form.
Chapter 7 sums up the main findings of each part of the book.
It is important to distinguish between intent and execution when trying to evaluate this book: the former is laudable, but I've found the latter lacking in many respects.
The thrust of the book is summarized on page 3: “Diachronic change of anaphoric systems is inherently interesting because it can provide evidence likely to inform our knowledge of those systems more widely, in particular where issues pertaining to Binding Theory are concerned”. This is yet another shove against the diachrony/synchrony barrier, a move that is becoming more and more common in recent years with the advent of usage-based approaches to grammar, to which the author subscribes overtly. Also, in line with Construction Grammar, the author intends to show that the evolution of clause-mate anaphor in French isn't one smooth change, but rather a jagged line in which different predicates evolve at different rates. Hence, a methodology based on corpora. Also crucial, finally, is the author's stance that languages are meant to be used and evolve because speakers do things with them, not because of imperfect learning. In other words, language change is an inherent property of language, not the side-effect of an external phenomenon.
That is the intent, and I couldn't be more sympathetic; but I was disappointed that the book didn't live up to those expectations.
First, the book has a patchwork quality that is quite puzzling to the reader; the central chapters (3 to 5) aren't tied together, as if they were investigating different phenomena, even though those phenomena involve the same markers. In particular, it is very surprising that the “soi” versus “lui”/ “elle” alternation and the simple versus reinforced one are kept distinct; one would think that the two phenomena have influenced each other, and if they didn't, that should still be verified. Similarly, the (not so short) analyses of French negation and compound past tenses in Chapter 6, just to illustrate the author's point, feel quite out of place.
Second, the author spends a lot of time on theoretical points, and much less on data. Although reviewing literature is an important step in any research, I couldn't quite understand why generative hypotheses are discussed at such length in Chapter 2, all the more as the author subscribes to Construction Grammar (the cognitive approach of van Hoek, 1995 isn't even mentioned). Chapter 4 is even more surprising, as the author develops his own hypothesis (i.e. “même” marks the predicate, not the pronoun) after discussing previous works, but as far as the history of French is concerned, there is virtually no result mentioned. Chapter 5, and especially Chapter 6, have the same feeling.
Admittedly, Chapter 3 is concerned almost exclusively with data, and the distribution of “soi” versus “lui”/ “elle” is even analyzed predicate by predicate. That is the most satisfying part of the book, or the one with solid results. More could (and, I believe, should) have been done in that direction; for instance, how is it that the distribution is influenced by the larger context, in particular, by the risk of ambiguity? Compare:
(3) Jean entra. Pierre parlait de lui/soi.
Jean came in. Pierre was talking about himself.
(4) Marie entra. Pierre parlait de lui/soi.
Marie came in. Pierre was talking about himself.
In (3), “lui” may refer to both “Jean” or “Pierre”; this is not the case in (4), under the assumption that no other male participant was previously mentioned, because “lui” cannot refer to a female participant, thus excluding “Marie” as a possible antecedent. Is “soi” more often used in contexts like (3) than in contexts like (4)?
The author links the use of “soi” to non-specificity, but only takes into account the specificity of the subject. It would have been interesting to consider aspect too:
(5) Pierre parle souvent de soi/lui.
Pierre often talks about himself.
(6) Pierre est en train de parler de soi/lui.
Pierre is talking about himself.
Is “soi” favored in (5) when compared to (6)? Perhaps there is no such tendancy, but I think it would have been worth investigating and much more illuminating than an advanced discussion of Binding Theory. Based on the fact that this is a usage-based-minded, corpora-driven study, I expected a fine analysis of usage as exemplified by attested utterances.
Finally, the editing of the book was careless. It is riddled with typos, many references are missing in the bibliography (e.g. Bouhours, 1676 and Vaugelas, 1647 on page 78, Brunot, 1966 on page 123, Kemmer, 1993 on page 139, Bhat, 2004 on page 140, and so on), and the publisher uses a font (Robert Slimbach's Minion Pro) without grooming it first. It is widely known (and visible) that at least this font’s apostrophe character needs grooming, especially with French, where, e.g., “l'usage” (‘the custom’) otherwise looks like “lusage”, with a floating accent. The author seems to have been aware of that, and therefore, introduced spurious spaces in the manuscript -- where sometimes the line was unfortunately broken (see example (32) on page 32). One may wonder whether anybody really proofread the book on the publisher's side, which is a very annoying feeling for which the author cannot be held responsible.
All in all, the book feels rushed; it would have been much stronger with a greater deal of detailed analysis rather than theorizing. This may of course be a matter of personal preference, but I strongly believe that corpus studies, and Construction Grammar on the theoretical side, have cut down hairsplitting hypotheses in favor of heavy data, and that such an approach is their major strength.
Due to its subject and its methodology, the book will be of interest to researchers working on French in either theoretical syntax or historical linguistics. It is probably too specialized to help students, but the results it contains can perfectly serve as examples in a course in historical syntax.
Chomsky, Noam. 1981. Lectures on Government and Binding. Dordrecht: Foris.
Detges, Ulrich and Richard Waltereit. 2002. Reanalysis vs. grammaticalization: A semantic-pragmatic account of functional change in grammar. Zeitschrift für Sprachwissenschaft 21. 151-195.
Reuland, Eric. 2005. Binding conditions: How are they derived? in S. Müller (ed.). Proceedings of the HPSG05 Conference. Stanford: CSLI.
Siemund, Peter. 2000. Intensifiers in English and German. A Comparison. London: Routledge.
van Hoek, Karen. 1995. Conceptual Reference Points: A Cognitive Grammar Account of Pronominal Anaphora Constraints. Language 71(2). 310-340.
ABOUT THE REVIEWER:
Paul Isambert holds a PhD from the University of Paris 3, France. He is currently working on grammaticalization of discourse markers and teaches at the University of Tours, France.