Morphologisches Minimum is a study of changes in verbal inflection (primarily the present indicative forms of haben and sein) and the forms of the personal pronouns in local dialects of High German between 1887, the date of Wenker’s Sprachatlas des deutschen Reichs, and the late twentieth century, when a number of major regional dialect atlas projects were completed covering much of the same area that Wenker had surveyed roughly a century earlier. Stefan Rabanus opens his monograph with a brief foreword that explains the basic rationale for the study and informs us that it is Rabanus’ Habilitationsschrift. This is followed by seven chapters, a bibliography, an appendix, and an index.

Chapter 1 introduces some basic concepts and lays out Rabanus’ general approach to language change and variation. Chapter 2 deals with descriptive morphological issues, discussing concepts such as syncretism, paradigms, and inflectional classes, as well as the grammatical categories that are relevant to this study: gender, case, number, and person. The chapter also explains what Rabanus means by Minimalsatz, a crucial morphosyntactic unit for his analysis that consists of a conjugated verb along with personal pronouns for all obligatory arguments. Chapter 3 discusses several approaches to morphological theory that are potentially relevant for the phenomena covered, focusing primarily on the work of German scholars. After a general introductory section on markedness, the chapter provides brief overviews of Natural Morphology, as developed especially by Mayerthaler, Dressler, and Wurzel, the Minimalist Morphology of Wunderlich and Fabri, and Optimality Theory. This last section also includes some comments on Distributed Morphology.

Chapter 4 provides a brief introduction to the relevant aspects of German verbal and pronominal morphology. Chapter 5 is devoted mainly to thorough descriptions of Rabanus’ data sources, which include the hand-drawn maps and original completed questionnaires of Georg Wenker’s Sprachatlas des Deutschen Reiches. Digitized versions of the questionnaires are now available online at http://www.diwa.info. The chapter also provides seven major regional dialect-atlas projects based on data collected in the last decades of the twentieth century. Rabanus explains that much of the material from these later projects had not yet been published when he was conducting his own study (through 2006), and that he therefore relied largely on unpublished sketch maps, lists of forms, and completed questionnaires. The chapter concludes with a short section on Rabanus’ own methodology.

Chapter 6 is the empirical heart of the book and, at 167 pages, is longer than the other six chapters combined. Each of the nine major sections deals with one region, describing and analyzing the paradigms of personal pronouns and the present indicative active forms of haben and sein (along with other verbs in some cases) found in that region. The regions are defined politically (Lower, Middle, and Upper Franconia, Bavarian Swabia, Alsace-Lorraine, etc.) rather than dialectally, so that distinct paradigms must be described for a number of different zones within each region. Following synchronic descriptions of the 1887 and the late twentieth-century systems, the discussion focuses on the diachronic correspondences. Any correspondence that is judged to be morphologically significant is described and analyzed, but, as the volume’s subtitle indicates, Rabanus is especially interested in changes that involve the loss of earlier paradigmatic distinctions or the emergence of new ones. Among such changes he devotes the most attention to cases where he sees a correlation between a change in the personal pronouns and one in verbal inflection. Rabanus interprets such correlations as evidence for his morphological-minimum principle, which he accounts for in terms of speakers’ desire or “need” to avoid ambiguity within the minimal sentence. Rabanus also recognizes a need to avoid excessive redundancy, giving us an upper limit (“obere Grenze,” p. 262) on the redundant marking of grammatical distinctions along with the lower limit (“untere Grenze”) enshrined in the morphological-minimum principle.

The final chapter (chapter 7) summarizes and elaborates on the key points made in chapter 6 concerning the roles of various factors in morphological change. The role of the avoidance of ambiguity (and redundancy) is again emphasized, but there are also sections on markedness, analogy and uniformity, frequency, ease of articulation, and language and dialect contact. The section on markedness is largely devoted to a discussion of the surprising fact that it is often the syntactically more marked form that wins out when distinctions are eliminated in the pronouns, as when accusative/dative forms like 1st plural uns and 2nd plural (Bavarian) enk replace the old nominatives wir/mir and es, respectively.

The great strength of this book lies in the enormous wealth of data that it presents and the often very original and effective ways in which Rabanus organizes and presents these data, especially in his tables and maps. Rabanus starts by categorizing the relatively raw data provided by the dialect atlas projects into morphosyntactically significant types and then shows how these types pattern geographically and how these configurations change over time. He makes his maps especially valuable and revealing by: 1) always presenting them in pairs, with one map showing the situation in Wenker’s 1887 survey and another showing the parallel situation in the late 20th century; 2) using symbols creatively and effectively to show combinations of morphological features, such as a verbal agreement ending and the corresponding nominative personal pronoun. Rabanus’ 102 tables and nine pairs of feature maps suggest patterns and correlations that would never be apparent from the dialect atlases themselves and that are likely to lead to major new insights into the factors that shape morphological variation and change. One minor quibble here: A complete list of tables and maps would greatly enhance the usefulness of the book.

Morphologisches Minimum is generally well organized, clearly written, and carefully edited and proofread. I found only the following typographical and similar errors: “Symbolierungen” for “Symbolisier-ungen” (p. 43); “buab-a” for “buam-a” (p. 46); “Ein der wichtiges Ergebnisse” (p. 91); “Allophonen” for “Allomorphen” (p. 133); “den den Infinitiv” for “den Infinitiv” (p. 133); “Personal” for “Person” (p. 151); “hab-n > hab-n, hab-n” for “hab-n > hab-n, ha-n” (p. 201); “ha auch vor Konsonant” for (presumably) “ha auch vor Vokal” (p. 206); “handelt es es sich” for “handelt es sich;” “eineindeutig” for “eindeutig” (p. 261).

One serious problem with the descriptive part of the book concerns null subjects (pro drop). Rabanus’ analyses of distinctions and ambiguities in the minimal sentence are based on the assumption that all the varieties under discussion are entirely non-pro-drop (compare pp. 67–69), that is, that every minimal sentence includes a subject pronoun. Null subjects, along with complementizer agreement, are in fact the norm, however, in the 2nd singular and plural throughout Bavarian, and in the 1st plural in Lower Bavaria (Weiß 1998, Fuß 2004). As I explain below, this is especially problematic for Rabanus’ lengthy discussion of the case syncretism involving the 2nd pluralpronoun enk(s) in parts of the North Bavarian dialect region.

The generally impressive success of Rabanus’ descriptive efforts is not, in my opinion, duplicated in his attempts to explain the patterns that he has uncovered. The author acknowledges that many of the changes he discusses have nothing to do with maintenance of a morphological minimum, and he is often undoubtedly on the right track when he accounts for these in terms of familiar factors such as analogy, uniformity, token frequency effects, etc. These factors are clearly of secondary interest to Rabanus, however, and the sections that deal with them are generally not as carefully researched or as well developed as other parts of Morphologisches Minimum.

Rabanus takes an extreme position in the debate over the proper place of teleological explanation in historical linguistics, arguing that the only truly explanatory accounts of language change are those based on final causes (goals), specifically on individual speakers’ desire for social success (p. 14). There may be sound philosophical arguments for defining “explanation” in this way, but it is not at all obvious that explanation in this narrow sense should be the main objective of historical linguistics. I would argue that Rabanus’ quest for explanations frequently causes him to overlook or downplay much more interesting aspects of the phenomena he is studying.

Rabanus focuses on aspects of grammatical change where it is relatively obvious how speakers’ intentions might play an important role. Specifically, where he sees a correlation between a paradigmatic distinction in verbal inflection and a lack of the corresponding distinction in the personal pronouns, or vice versa, he argues that this correlation must be attributed to the morphological-minimum principle and to speakers’ desire or need to avoid ambiguity and excessive redundancy. (This explanation is combined with the—unexplained—fact that German has not followed the path taken by English and some other Germanic languages of relying increasingly on word order to represent case distinctions.) In chapter 6, Rabanus identifies a number of correlations that he attributes to the morphological-minimum principle. In chapter 7, he highlights three of these, which he regards as providing especially clear evidence for the roles of avoidance of ambiguity and avoidance of excessive redundancy in language change (p. 262).

The first of these concerns the use of the 1st plural pronominal form uns for nominative (in addition to accusative and dative) and the 1st plural verb ending -mr in an area in the northwest of Bavarian Swabia (pp. 172–178, 189–192, 262–266). Rabanus shows that in 1887, 22 locations reported using nominative uns but not -mr, instead using -nt or -et as the agreement ending for all persons in the plural (as do areas to the south which do not have nominative uns). This situation entails a potential ambiguity since these dialects, like most other varieties of German, do not distinguish between nominative and accusative in the 3rd plural pronouns either. Thus, a sentence such as uns seet dia could either mean ‘we see them’ or ‘they see us’. In 1989, this potential ambiguity has been eliminated everywhere. All locations with nominative uns now also have -mr, which Rabanus regards as clear evidence of the role of the morphological minimum in motivating change.

Rabanus also includes another change on the combination maps he uses to show the spread of -mr (maps 13 and 14, pp. 176–177): The 2nd plural pronoun ui, which occurs throughout most of Bavarian Swabia in the accusative and dative only, is extended to the nominative in an area considerably larger than that with -mr. He acknowledges that this non-case-marked pronoun for the 2nd plural creates a new ambiguity for which the verbal inflection of the area does not in any way compensate (pp. 192, 265). I have more to say about his attempts to explain the spread of non-case-marked ui below, but here the point is simply that on balance the data presented on maps 13 and 14 look like anything but support for the morphological-minimum principle. By my count, the changes shown indicate that between 1887 and 1989 five locations pick up both ambiguities (1st and 2nd plural), 27 show a net increase of one ambiguity, and 38 have no net change, compared with just five showing a net loss of one ambiguity.

Rababus’ second example involves the spread of the non-case-marked 2nd plural pronoun enk(s) in the northwestern Upper Palatinate and adjoining areas of Middle Franconia (pp. 223–230, 266–268). Rabanus shows that the geographic distribution of non-case-marked enk(s) in 1887 and its further spread in 1999 correlate strongly with the distinctive 2nd plural verb ending -ts. He particularly emphasizes that non-case-marked enk(s) does not spread to a single location with -t as the 2nd plural verb ending and argues that this correlation demands a “functional” (p. 229) explanation in terms of avoidance of ambiguity: In dialects with -t as the 2nd plural verb ending, the 2nd plural form is identical to the 3rd singular form for many verbs. Thus, a hypothetical sentence such as enk(s) baes-t dei, with non-case-marked enk(s), would be ambiguous between ‘you(pl) bite her’ and ‘she bites you(pl)’. This ambiguity is avoided in dialects with -t as 2nd plural ending by distinguishing between nominative and accusative/dative in the 2nd plural pronouns, while all dialects that do not distinguish case in these pronouns have the distinctive 2nd plural ending.

A major problem with this line of argumentation becomes apparent when we compare the geographic correlation between case syncretism and distinctive verb endings in the 2nd plural in this region with that found in Bavarian Swabia, for which Rabanus’ map pair 15 (p. 187) reveals essentially the opposite of the correlation found in and around the Upper Palatinate. Here, it is on the Bavarian side of the dialect border, with the distinctive -ts 2nd plural verb ending, that we see no sign of case syncretism in the 2nd plural pronouns. In contrast, on the Swabian side (in Rabanus’ zone XVIII), where the 2nd plural verb ending is generally -nt, we find the non-case-marked pronoun ui, as shown on map 14 (p. 177). The -nt ending is distinct from 3rd singular -t but is identical to the 3rd plural, which would, in fact, lead to even more ambiguities since many verbs undergo distinctive stem alternations in the 3rd singular form. This comparison of the correlations in the two regions suggests that they may have nothing to do with avoidance of ambiguity, and that we need to consider other kinds of explanations.

For enk(s) and -ts in the Upper Palatinate, two possibilities immediately come to mind, either one of which could, in my opinion, account for the observed correlation more plausibly than avoidance of ambiguity. The first is a straightforward correlation that Rabanus does not mention: Between 1887 and 1999, non-case-marked enk(s) spreads only to dialects that already had enk(s) in either the nominative or (much more often) the dative/accusative in 1887. These dialects are (for well-known historical reasons) generally the same ones that have -ts, but one could reasonably argue that it is actually the presence of enk(s) in the paradigm rather than any direct connection with the distinctive verb ending, that makes these dialects candidates for non-case-marked enk(s).

The second factor, already mentioned above, is that it is precisely the dialects with the -ts ending that normally have null subjects in the 2nd person (along with complementizer agreement in subordinate clauses). Since the unmarked 2nd plural subject pronoun throughout Bavarian is actually Ø, it is quite misleading to treat the use of enk(s) for (emphatic) nominative as fully comparable to the other instances of case syncretism discussed in Morphologisches Minimum. One consequence of Bavarian pro-drop would be that the amount of evidence available to learners for the shape of the 2nd plural subject pronoun, which, as Rabanus points out (pp. 289–290), is quite low in frequency to begin with, would be severely limited. It is thus not surprising that speakers of these dialects would often simply fail to fully acquire these subject forms and would resort to using another 2nd plural pronominal form instead.

Rabanus’ final example of a correlation between loss of one distinction and emergence or maintenance of another involves the 2nd plural in Kreis Lörrach and a small adjacent section of Waldshut, in the southwestern corner of Germany. Here we find some dialects with a distinctive ending in the 2nd plural while others have the same ending throughout the plural. The nominative form of the 2nd plural pronoun is ihr in some locations and dir in others, the latter form being homo-phonous with the 2nd singular dative pronoun. Rabanus claims: “Hier haben der Wandel vom verbalen Einheitsplural zum Zweiformenplural des Typs DED und die Ausbildung eines Kasus- und Numerus-synkretismus beim Pronomen der 2. Person dieselbe areale Extension […]” (p. 268, compare p. 153), and more specifically: “Der pronominale Synkretismus der 2. Person in der Form dir ist […] dadurch funktional abgesichert, dass der durch diesen Synkretismus verursachte Verlust an morphologischer Kodierung durch die Ausbildung des Zweiformen-plurals DED kompensiert wird.”Footnote ²

Assuming that this correlation between distinctive 2nd plural verb endings (in -t) and the 2nd plural nominative pronoun dir is real, there are once again highly plausible explanations for it that have nothing to do with avoidance of ambiguity. The emergence of -n/-nt/-n or -Ø/-t /-Ø paradigms in the verb plurals where the 2nd plural nominative pronoun begins with d- looks like a straightforward case of reanalysis (resegmen-tation), exactly parallel to the well-known change that gave us -st in place of earlier -s for the 2nd singular ending in OHG (pp. 116–117). This reanalysis is essentially the reverse of the one that gave rise to 2nd plural nominative dir (< earlier ihr) in the first place (p. 121). Rabanus’ map pair 9 (p. 154) shows that this original ihr > dir reanalysis is still spreading in the southwest corner of area XIII (LÖ 21, 25, 26, 27, 29, 30) and perhaps also further east(LÖ 2, WT 5, WT 21). Here, the precon-dition for the reanalysis is a 2nd plural verbal ending in -t. Where the -t has been lost, as in the section of Kreis Freiburg shown on Rabanus’ maps, we see, as expected, no further progress of ihr > dir, instead finding this change being reversed in the few locations where it had occurred (FR 33, 34, 45) under the influence of neighboring dialects that had retained ihr. As in the previous case, evidence from Bavarian Swabia strengthens the suspicion that the apparent correlation between the ihr > dir change and the distinctive 2nd plural verb endings has nothing to do with avoidance of ambiguity. The author tells us that dir is robust and spreading in Bavarian Swabia not only in the far north (his zone XXIII), with its -e/-t/-e plural paradigm, but also throughout the southern half of the region (his zones XIX and XX), where the verb ending is -et through-out the plural (pp. 170–171, 192–193).

A very general problem with Rabanus’ strategy—building an em-pirical case for the morphological-minimum principle around instances where 1887 violations of the principle are later repaired—is that it simply begs the question of how and why the violations arose in the first place. Rabanus says very little about the factors motivating most pre-1887 developments, but he addresses the question indirectly in his above-mentioned discussions of the spread of the non-case-marked 2nd plural pronoun ui in northwestern Bavarian Swabia and the prevalence of the 2nd plural nominative pronoun dir in the southern half of the same region. Here, he argues that other factors can overrule the morphological-minimum principle. In the case of ui, he proposes that the relevant factors are “Analogie und Uniformität: [das] Bedürfnis nach einheit-lichem Aufbau zusammengehöriger Teilparadigmen […]” [Analogy and uniformity: [the] need for consistent structure of corresponding sub-paradigms] (p. 265, compare p. 193). Rabanus thus regards the morpho-logical-minimum principle as a violable constraint and states his intention to model its interaction with other constraints, such as uniformity, in an Optimality Theoretic framework in future work.

Perhaps this promised model of constraint interaction will show that the many apparent counterexamples to the morphological-minimum principle are not so problematic after all, but in the absence of such a model it is hard to know what to make of the principle. Avoidance of ambiguity undoubtedly does play some role in language change, but it is clearly far from being the whole story, and the fact that it can easily be “explained” in terms of speakers’ intentions and goals does not by itself make it a particularly interesting or important part of the story.

This brings us back to the book’s great strength. It is the tremendous wealth of data Rabanus provides and the detailed, accurate, and creative ways in which he presents those data that allow the reader to evaluate his theoretical proposals for him/herself. This is a rare quality in a project dealing with such enormously complex linguistic phenomena, and all criticisms of Rabanus’ theoretical arguments pale in comparison with this impressive achievement. Morphologisches Minimum is one of the best examples I have seen of what a theoretically minded linguist can do with the incomparable empirical treasures produced by the great German dialect atlas projects of the nineteenth and twentieth centuries. It deserves to be widely read, discussed, and critiqued by all scholars who recognize the value of basing our theories of language change and variation on the most solid empirical foundations available.