LINGUIST List 15.284

Mon Jan 26 2004

Review: Typology/Morphology: Cysouw (2003)

Editor for this issue: Naomi Ogasawara <naomilinguistlist.org>


What follows is a review or discussion note contributed to our Book Discussion Forum. We expect discussions to be informal and interactive; and the author of the book discussed is cordially invited to join in.

If you are interested in leading a book discussion, look for books announced on LINGUIST as "available for review." Then contact Sheila Dooley Collberg at collberglinguistlist.org.

Directory

  • Wolfgang Schulze, The Paradigmatic Structure of Person Marking

    Message 1: The Paradigmatic Structure of Person Marking

    Date: Sun, 25 Jan 2004 23:59:17 -0500 (EST)
    From: Wolfgang Schulze <W.Schulzelrz.uni-muenchen.de>
    Subject: The Paradigmatic Structure of Person Marking


    Cysouw, Michael (2003) The Paradigmatic Structure of Person Marking, Oxford University Press, Oxford Studies in Typology and Linguistic Theory.

    Announced at http://linguistlist.org/issues/14/14-1969.html

    Wolfgang Schulze, University of Munich

    INTRODUCTION

    One of the perhaps most uncontroversial claims related to universal aspects of human language concerns the concept of personhood. In linguistic work, this conceptual layer often is taken for granted and escaped from further elaboration, especially in descriptive work. On the other hand, personhood has been a crucial issue especially in language philosophy, yet rarely reflected in standard linguistic treatises. Not surprisingly, Paul Forchheimer's famous doctoral dissertation (Forchheimer 1953) has remained the only comprehensive survey on 'personal pronouns' for fifty years. Admittedly, a number of individual papers and allusions to the issue in studies related to typology and semantics have appeared on the linguistic market deepening the insights in the linguistic expression of personhood. Nevertheless, the author of the book under review can only be praised for having undertaken the enterprise not just to revise Forchheimer's approach, but to present a typology of personhood that is based on contemporary cross-linguistic methodology and that exploits both the specialized literature published since Forchheimer 1953 as well as proposals related to typological generalizations.

    To say it from the beginning: What we have now at hands is perhaps the best survey on the paradigmatic organisation of linguistic personhood ever compiled. Micheal Cysouw's 'The Paradigmatic Structure of Person Marking' (henceforth PSPM) represents a revised version of his 2001 University of Nijmegen doctoral dissertation and we can only thank Oxford UP for having accepted the book to be published in its 'Oxford Studies in typology and Linguistic Theory' series. The inclusion of PSPM into this series not only guarantees the attention of the world- wide linguistic audience but also conditions that Cysouw's work is packed into a well-done and appealing format. Hence, we can expect that PSPM will soon become a standard book of reference for issue related to the morphology of personhood that by far exceeds the quality of its predecessor, namely Forchheimer 1953.

    Nevertheless, PSPM cannot be uncontroversial. The reader should constantly recall the title of the volume that focuses on 'paradigmatic structure'. In other words, it deals with aspects of form that are related to functions. This classical form-to-function (here: paradigm- to-function) approach necessitates certain deductive claims on categorial and semantic issues used as a 'tertium comparationis' in Cysouw's formal typology. However, Cysouw avoids pathways that would start with a general discussion of categorial aspects of personhood as present for instance in the tradition of language philosophy, in language sociology, and linguistic psychology. Perhaps, this reluctance to design a 'semantic' template of personhood is due to the fact that even in very recent approaches to the semantics of linguistic units as expressed for instance in the framework of Cognitive Semantics, the question of personhood rarely exceeds very general statements related to the function of 'person markers' in a speech act. In PSPM, the author devotes just three pages (pp.5-7) to discuss some semantic issues of personhood. Not surprisingly, he makes reference especially to those authors who relate personhood to its role in a turn-taking cluster of speech acts (especially Goffman 1979, Levinson 1988). Accordingly, his definition of 'person marker' reads as follows: ''They have to be a shifter, specialized for that function, and used for reference to speech act participants'' (p.5). Much can be said about this delimitation of person markers, which clearly shows that Cysouw is not too much interested in the underlying semantics of such markers. For instance, it soon comes clear that the definition does not necessarily hold for so-called non-Speech Act Participants (termed 'other' by Cysouw): This category (if ever it is a linguistic category at all) is not marked for shifter functions, nor do the corresponding forms always specialize in the given function (e.g. demonstrative pronouns used as third person pronouns). In addition, they do not (as their name tells) make reference to speech act participants as such. In fact, a semantic-based typology of personhood would probably have to start from a rather different categorial setting, which would transgress most of the constraints or delimitations set up by the author. Most crucially, it would distinguish a cognitive layer of personhood from its pragmatization in discourse (e.g. in the sense of Mead 1934, Mauss 1938, Schulze 1998:575-601).

    It thus comes clear that the reader will appreciate PSPM especially if (s)he has adopted the author's form-to-function approach. A different approach, e.g. based on Cognitive Typology, would probably have led to an alternative design of the paradigmatic embedding of linguistic forms related to the concept of 'person' and of the paradigm internal dynamics (see below). Perhaps it is one of the few shortcomings of PSPM that it does not draw the reader's attention to this fact.

    PSPM is an extremely rich book, full of data and stimulating observations. It is out of question that the author had developed an invaluable tool to handle the paradigmatic structures of personhood. His cross-linguistic approach is based on the analysis of the relevant paradigms in more than four hundred languages and hence represents one of the broadest cross-linguistic studies ever prepared. In this review, it is impossible to account for all types of paradigmatic variation as they are elaborated in PSPM. The reader will greatly enjoy both the presentation of these paradigms: Cysouw uses a very helpful schematic representation which allows the reader to constantly refer to the general paradigmatic space described by the author and to locate the given data in this space. The reader will also profit from the careful presentation of the data which are constantly checked against their sources and thus can serve as a reliable data base for further studies.

    Instead of detailing out the universe of paradigmatic variation and dynamics, I will briefly describe the overall scheme of the book under review before turning to some general remarks on the approach advocated for by the author.

    OVERVIEW

    PSPM comprises xiv+375 pages, divided into four major parts, which again are enclosed by an 'Introduction' (pp.1-35) and a 'Finale: Summary and Prospects' (pp. 295-321). In order to help the reader to easily retrieve information, the book gives a list of languages according to their genetic/geographical distribution, and three indices (names, languages, and subjects). The nature of the book conditions that the list of references is of considerable size (roughly some 600 entries). The fact that the author has consulted primary sources as much as possible, illustrates the tantalizing work Cysouw has undertaken. Nevertheless, it must be added that certain relevant pieces of literature are missing, such as Russel 1940, Mead 1934, Mauss 1938, Anscombe 1981, Kantor 1952, L�vi-Strauss 1962, Myrkin 1964, or Majtinskaya 1969.

    The introductory chapter nicely outlines the scope and objectives of PSPM. Most importantly, Cysouw makes clear that once the set of personal markers have been delimited, they have to be analysed in terms of the paradigms they establish. The 'content' of such paradigms may exceed or go behind of what we know from Standard Average European, to use a nevertheless problematic term. Just in the beginning of the Introduction, Cysouw acquaints the reader with the famous passage from Domingo de Santo Tom�s' 'Gramm�tica o arte de la lengua general de los �ndios de los Reynos del Peru' (1560) that lays the ground for the well-known distinction between an 'inclusive' and an 'exclusive' 1pl. We have to thank Cysouw for having made available again this important passage (p.2) both in its original and in translation.

    The Introduction also discusses the question of defining the paradigmatic space of personhood, addresses methodological issues and gives a brief report on previous cross-linguistic investigations on the given topic.

    Chapter 2 and Chapter 3 discuss basic aspects of person marking. In Chapter 2, Cysouw refers to the ''The Marking of Singular Participants'',y"cely entitled ''One among the Crowd'' (p.41). Here, the author concentrates on questions of homophony, suggesting that a) such a homophony is comparatively rare, and that b) it is restricted to inflectional patterns. He argues for nine 'logical' types of homophony, eight of which he is able to back up. The ninth type is marked by second person = zero, 1/3 is marked, hence we have to deal with the opposite of e.g. the German type _ging_ (1/3) vs. _gingst_ (2) 'went'. However, if we look at admittedly rather specialized paradigms such as the cluster imperative-hortative, we may add many such examples, compare Udi (Lezgian) _campazu_ (1), _campa_ (2), _campane_ (3) 'write' (note that here, only markedness is used as a criterion, not the form of the marker itself). In addition, Cysouw describes the degree of variation that can be found with respect to zero-markers.

    In Chapter 3, the view is extended to 'group marking'. Here, Cysouw suggests a very helpful redefinition of the notion of plurality. He arrives at a very important point, saying: ''First, the multiple persons or objects have to be in the same predicative role. Second, the morpheme has to be unmarked as to the specific amount of elements. Finally, the morpheme should not include the regular reference for any singular person or object'' (p.67). Undoubtedly, this chapter represents one of the many highlights of the book. It carefully introduces the reader to the world of pronominal 'plurality'.

    Chapter 4 and 5 extend the discussion prepared in the preceding chapters to a typology of paradigmatic structuring. Cysouw develops his typology with the help of mnemotechnically useful terms that stand for specific 'types' (e.g. Latin-type, Sinhalese-type etc.). Each of these types serves as a template to discuss variations within this type. Again, he distinguishes non-homophonous split-types from those which show some kinds of homophony. A further distinction concerns the presence of the inclusive/exclusive dichotomy. Cysouw arrives at a total of sixty-three paradigmatic structures (p.165) and advocates against the assumption that most of these types represent corrupt or extended version of the SAE type marked for the typical six-way paradigm. Nevertheless, his findings allow him to observe eight common types, five semi-common types, whereas the bulk of paradigmatic variation is characterized as 'rare'.

    Discussing 'compound forms' in Chapter 5, the author clearly argues in favour of an 'incorporation' strategy. Compound forms are those that either cumulate different persons (e.g. 'we+you') or render a given reading of a pronominal form more explicit (e.g. the Russian inclusive _my s toboy_ ('we [that is] with you', a striking analogy can be found in Inuktitut, see PSPM, p. 183)). Hence, Cysouw distinguishes a cumulative reading from an incorporative reading and arrives at the following conclusion: ''The referential value of the compound pronouns either builds categories that are well known (...) or has identical reference as to the non-singular simplex pronouns'' (p.184).

    Chapter 6 and 7 turn to 'true' number forms. Having eliminated the concept of 'plurality' from the descriptive frame for quantitative reference with person markers, Cysouw correctly assumes that person markers are sensitive for number only if they refer to a concrete number of persons/objects. The author proposes the category of 'restricted groups' to denote this number layer. Restricted groups usually turn up as duals or trials. The (basically graphic) metalanguage developed by Cysouw is a powerful tool to account for variations in number. Nevertheless, it should be added that the glosses are slightly irritating, because the 'non-restricted group' section lacks such a corresponding glossing that would indicate the kind of reference towards person. Perhaps, it would have made sense to add glosses such as '1+2x' (= EGO + unspecified number of TUs), '2x' (= unspecified number of TUs), or '1+3x' (= EGO + unspecified number of non-Speech Act Participants' etc.).

    Chapter 6 introduces the reader to the world of restricted number. Here, the author extensively discusses the problem of interpreting the inclusive, which Cysouw views as an ambiguous category (wavering between sg, du, and pl). Chapter 7 turns to ''the diversity of restricted groups: a survey of dual person marking''. Both duals with and without an inclusive/exclusive distinction are discussed, again concentrating on the types of homophony observed especially with bound pronouns (in sum thirty three paradigms).

    Chapter 8 turns to the diachronic dimension. Note that Cysouw uses the term 'crypto-diachronic method', which he describes as follows: ''The method (...) is not a historical comparison, but a typological comparison that starts from the broad typological generalization and tunes into the fine-grained differences with a genetic group'' (p.247). The author sets up four conditions 'cognate' paradigms have to meet in order to be taken into consideration (p.248). Curiously enough, 'regular sound correspondences' are not among these criteria, although diachronic shifts present with the formal expression of person markers are crucial for both determining a cognate set and the conditions of change. Instead, Cysouw concentrates on questions of homophony and degree of explicitness in order to describe paradigmatic change. In other words: Paradigms are taken as some kind of 'gestalt' that tend to change with respect to their properties over time. The fact that Cysouw neglects the dynamics of sound changes renders it difficult to understand what he means by ''phonologically closely related'' (p.268). How does the author decide to decide that two forms are related according to this criterion?

    For instance, the southern German dialects usually called 'Allemanic' show a third person plural 'bound morpheme' (present tense) _-et_ instead of _-en_. This difference accounts for a varying pattern of homophony (Allemanic 3sg+2pl+3pl / 2pl vs. Standard German 1pl+3pl / 3sg+2pl). If we would not know that the 3pl stems from _-ent_ that regularly developed into _-et_ in Allemanic, but to _-en_ in other dialects of German, we would perhaps assume that the paradigms are not related (admittedly, I cannot say whether Cysouw would interpret the 'correspondence' -_t_ ~ _-n_ as being 'phonologically close related'). Or: How can we decide that German _-st_ (2sg) is related to say Latin _-s_ (2sg), but not to _-t_ (3sg), if we would not know that _-st_ is derived from *-s + *thu (you:sg)?

    These examples may be trivial from an Indoeuropean point of view. Nevertheless, I would like to stress that the criterion mentioned above should be taken with great caution because especially with languages that lack a scientifically elaborated diachronic, similarity in form may turn out as 'false morphological friends'. Or vice versa: Formally non-similar forms may be based on the same morpheme (recall the Armenian nominal plural _-k`_ which corresponds to say the Latin -s- Plural). On the other hand, formal 'similarity' is sometimes used to construe rather problematic assumptions on the categorial affinities. For instance, p. 272 Cysouw starts with the Nabak (a Finisterre-Huan language from Papua) third person (group) _ekngen_ and derives it from the 3sg _ek_ saying ''[t]he singular morpheme (...) is compounded with the second person non-singular [_in_] morphemes (sic!) to form third person non-singular forms''. It is indeed difficult to understand how this compounding should yield a third person non-plural. The homophonous 2/3 non-singular _gin_ of closely related Wantoat does not help very much, because here the third person singular is _an_. Hence, we would have first to prove that Nabak _ek_ and Wantoat _an_ are related. Only if this correspondence can be safely described with the help of sound laws, we may go on and try to analyse forms like Nabak _ekngen_).

    Nevertheless, it is out of question that Cysouw arrives at a very compelling picture of paradigmatic dynamics. Personally, I doubt whether we can go so far to design a 'cognitive map of interconnected paradigmatic structures' based on the criteria set up by the author (p.268). Still, Cysouw's approach can surely help to better understand the dynamics of paradigmatic change.

    Chapter 9 extends the diachronic perspective to dual forms. The highly illuminating examples helps him to refine the generalizations of paradigmatic dynamics made in the previous chapter. He again refers to the two hierarchies suggested before (Explicitness Hierarchy and Horizontal Homophony Hierarchy), to which he adds the Dual Explicitness Hierarchy. A 'cognitive map of paradigmatic structure' nicely summarizes the options of paradigmatic change. Here, another word of caution seems appropriate: It is rather modern to refer to 'cognitive maps' in order to account for semantic or categorial correlations. However, each such 'cognitive map' should always be embedded into a more general theory that explains how paradigms are represented in cognition (if ever they are). Naturally, it is rather attractive and seemingly self-evident that the categories of personhood reflect a conceptual layer that organizes the inter-individual discourse. However, we may likewise assume that personhood is not a cognitive 'category' at all, but a side-effect of other cognitive (here: perceptual) 'mechanisms' (such as figure-ground parsing, empathy, distribution of knowledge among individuals and related presuppositions). In other words: What is at need is a cognitive theory of paradigms that goes beyond the standard assumptions of semantic and formal correlations. In the final chapter of PSPM, Cysouw first summarizes his central observations. He then tries to approach a 'theory of person marking' before turning to 'prospects'. Here, he turns the reader's attention to the correlation of 'independent' vs. 'inflectional' person marking. Accordingly, there is ''a correlation between less explicitness in the paradigm (meaning more horizontal homophony) and more inflectionally marked pronominal paradigms'' (p.313). In addition, Cysouw considers the crucial point of 'asymmetry of affixation' (segmenting a paradigm into prefixing and suffixing strategies). Here, the authors arrives at an interesting observation: ''[T]here is a correlation between the size of the paradigm and the affixial status. (...) [T]he smaller paradigms are more often prefixes and the larger paradigms are more often suffixes'' (p.316). Finally, Cysouw briefly alludes to the question of gender marking. Here, I would have appreciated a more concise treatment especially of the question whether gender really is a ''curious linguistic phenomenon'' (p.319) when present with the first singular. We should recall that grammatical descriptions rarely tell us whether an informant has been a woman or a man. Keeping in mind that in many societies, women are forbidden to interact with 'strangers', we may doubt that the description of personal paradigms sufficiently considers the language of women. In other words, gender studies are not a central issue of language typology. We have to assume that both the group- internal language of women as well as their self-reference in the 'male world' (in terms of social deixis) may considerably differ from what we can find in standard grammars.

    SOME CONCLUDING REMARKS

    It comes clear that the more 'deviations' from SAE-typical paradigms become, the more a well-defined delimitation of Person markers becomes necessary. Here, Cysouw argues that an analysis of personal paradigms must not be restricted to 'free' pronominal forms, but has to take into consideration 'bound' forms, too, as they may show up in term of inflection. For both he observes: ''Diachronically, person markers do not behave differently from other linguistic elements. They grammaticalize from independent nouns into person markers. Also independent pronouns grammaticalize into inflectional person markers'' (p.5). These two claims, however, cannot be left without comment: First, 'if' person markers may stem from 'nouns', Cysouw violates his own criterion of functional specialization at least from a diachronic point of view: it is a well-known fact that grammaticalization hardly ever happens in terms of function 'hopping', say from 'referential' to 'shifter'. Rather, we have to refer to a scenario (best described in terms of grammaticalization chains) that allow the (intermediate) co- existence of two or more functional domains in one and the same form, be it from a diatopic or a diastratic point of view. Hence, 'if' Cysouw's assumption about the origin of Person Markers is right, we have to assume that there must have been a time, when such markers served for more than just to indicate aspects of personhood (Cysouw himself draws the reader's attention this fact (p.13)). In order to save the case, the analysis has to be confined to paradigms at the 'end' of the grammaticalization chain. Doing so means to deprive oneself of the possibility to explain the make-up of a paradigm from a diachronic perspective, which again does not make sense, if one subscribes to the fact that paradigms of personhood belong to the universals of language (in what shape so ever).

    In addition, the criterion of functional 'specialization' presupposes that the category of person is an 'autonomous' category in language. However, we may likewise assume that 'person' results from the blending of different categorial or functional layers (see Mauss 1938). Both 'pronouns' and 'inflectional person markers' are always embedded into both constructional 'frames' and the general Speech Act typology. For instance, a first person singular is prototypically linked to assertions, whereas the second person singular is linked to modal constructions such as imperatives or questions. It may well be that a second person is - in its form - conditioned by the 'grammaticalization' of interrogative strategies (a nice example is German _-st_ (2sg) already referred to above: The clitization of _-t-_ < *_thu_ etc. ('you:sg') can only be understood, if we start from an interrogative pattern, which is marked by a verb-initial construction in polar yes/no-questions. From this, we may conclude that any personal marker is likely to encode (or to be semantically conditioned by) more than just the category of personhood. In other words: The delimitation as proposed by Cysouw serves heuristic purposes rather than 'cognitive reality'. For instance, the author counts the Lak (East Caucasian) clitics (not suffixes!) _-ra_ (1/2sg), _-ri_ (3), _-ru_ (1/2 non-sg) as person markers (p.127), although their primary function is that of a so-called bipolar focus that has a side-effect as for the category speech relevant (_-ra_ / _-ru_) vs. non-relevant (_-ri_). In fact, none of the three markers are person markers at all.

    Furthermore, Cysouw's claim quoted above alludes to popular assumptions on grammaticalization patterns that are perhaps to strongly oriented towards the path 'noun -> non-noun'. If we browse through the diachrony of those personal pronouns the history of which can be (more or less) safely described, we soon realize that it is extremely difficult to postulate with certainty the original nouniness of such pronouns. This difficulty is related to the observation that (singular) personal pronouns often belong to the most stable forms of a language from a diachronic point of view, both with respect to form and to function. Hence, when reconstructing a given pronoun, we often arrive at just another form of the pronoun, but not at something like a noun. On the other hand, Cysouw importantly neglects other hypotheses on the origin of personal pronouns that relate them to the set of demonstratives or more general to deictic terms (a nice example is Liebert's approach to Indoeuropean pronouns, see Liebert 1957, also confer Majtinskaya 1968, Schmidt 1978, Schmidt 1994).

    Finally, the quote also refers to what can be cautiously call a 'linguistic myth'. Accordingly, 'bound' morphemes used to subcategorize personhood are often regarded as grammaticalized versions of the corresponding free pronouns, whether or not these have been retained in a given language. Although Cysouw only states that pronouns may turn up as bound morphemes (which is correct), it is important to note that he does not tell whether other grammaticalization paths are possible, too (such as focus and other deictic markers, specialized copular constructions etc.) and which impact such paths may have had on the organization of a paradigm.

    As has been said above, the architecture of personal paradigms represents the core issue of PSPM. In order to render his data comparable, Cysouw has to make a number of further delimitations some of which are crucial for his arguments. Most importantly, he neglects functional clusters that involve personhood. Rather, he tends to split off such clusters and to describe the architecture for each paradigm separately. The main reason for doing is to guarantee cross-linguistic comparability: ''The result of this approach is an insight into the paradigmatic structure of person marking. Only indirectly will this help us to understand the functioning of a whole language'' (p.10). I am not quite sure whether the perspective taken by Cysouw is well-chosen. It disregards the possibility that paradigmatic structures are motivated by their co-paradigmatic environment. An example is already given above. Let me briefly illustrate this point with the help of another example: In Archi, another East Caucasian language, the first person singular _zon_ is embedded into a standard paradigm opposing sg to non-sg pronouns, if we look at the absolutive case. However, this pronoun is the only one which knows a distinct ergative form (_z�ri_). On the other hand, the second person (sg/non-sg) is the only categorial entity which does not reflect the noun class of its possessum if used as a possessor. Obviously, it is the specific conception of agenthood and possessorship in Archi that accounts for these paradigmatic 'split' types. In the Upper Andi variety of Andi, again an East Caucasian language, a slightly analogous split is found: Here, the 1sg _din_ has a distinct ergative form _den_ in the language of women, whereas men use _din_ for both absolutive and ergative. Curiously enough, the Keleb dialect of Avar, a language (distantly) related to Andi, turns the paradigm around: In this language, the 1sg ergative is present in the language of men, but not in the language of women. In Chechen, again an East Caucasian language, the emphatic-reflexive variants of the personal pronouns do not distinguish between 2pl and 3pl (_s^�s^_), whereas the non-emphatic variants do. In addition, the inclusive _vay_ (borrowed from an Indo-European language) is the only pronoun that does not know an ergative case.

    These examples illustrate that it would perhaps have made more sense to first construe the cognitive space of personhood in the individual languages before starting the comparison. This concerns both the referential behavior of 'pronouns' and the relational behavior of agreement forms (e.g. tense/aspect/mood; diathesis etc.). In other words: It seems doubtful that generalizations resulting from the cross- linguistic comparison of individuated paradigms will tell us more than just what is possible in language. Cysouw himself is well aware of this problem. Nevertheless, he argues: ''[A] paradigm is a set of linguistic elements that occur in the same syntagmatic place in the structure of a language'' (p.8). In order to also reflect TAM-categories etc., he should have added: ''and that belong to the 'same' superordinate paradigm''. However, contemporary grammar theories that belong to the camp of e.g. Cognitive Linguistics will likely challenge this view. Cysouw assumes that a person marker acquires its function only because it is embedded into 'its' paradigm that includes 'other' person(s). However, it is rather likely that the same holds for the syntactic and temporal-spatial embedding of such markers (Cysouw himself draws the reader's attention to this point (p.49)). In other words, a 1sg is a 'first singular' (EGO) also because it is embedded into the paradigm of agenthood, a 2sg is a second singular (TU) also because it is embedded into the paradigm of modality (question/command etc., see above). Accordingly, much depends from which dimension is referred to when describing 'paradigms'. The way Cysouw has chosen in PSPM is rather traditional and perhaps too strongly focused on the notion of 'speech act participation'.

    Finally, let me briefly turn to the data presented by the author. He writes: ''Every delimitation proposed for a cross-linguistic study is bound to encounter exceptions and problematic cases when confronted with the actual linguistic variation'' (p.19). Nevertheless, it can safely be said that Cysouw's presentation of the data and the choice of his sample exhibits a highly learnt approach that guarantees the high quality of his analysis. Unfortunately, his sample is somewhat biased by the evident neglect of Russian sources. This fact conditions that the domain of Turkic and Mongolian languages is strongly underrepresented. Modern Iranian languages are lacking completely, although their paradigms add crucial information. Another 'laboratory' of personhood, namely the East Caucasian languages are quoted (except for Hunzib) from second-hand sources only, some of them rather dubious or at least too superficial (see Schulze 1999 and Schulze 2003 for some details on East Caucasian pronouns). Hopefully, the deplorable fact that Russian sources are rarely respected in cross-linguistic comparison will soon face revision.

    SUMMARY

    In sum, PSPM represents a courageous and highly innovative approach to the paradigmatic architecture of personhood from a cross-linguistic perspective. Contrary to what Cysouw says, no parts of the book are boring or redundant. The author has developed a highly stimulating way of handling and presenting cross-linguistic data, which helps the reader to safely navigate through the world of linguistic variation. (S)he is well-equipped with a huge amount of data and a methodological 'compass' guaranteeing that (s)he never loses orientation. I am not quite sure whether the 'ship' Cysouw invites us to embark will bring us to the final destination, namely to a coherent 'theory of personhood in language'. May well be that Cysouw's journey through the ocean of person markers is comparable to Christopher Columbus' journeys that reached the Caribic islands but not (yet) the mainland. In other words: PSPM is an important step towards these mainland, which will be probably not reached without considering Cysouw's impressive work.

    REFERENCES

    Anscombe, G. E. M. 1981. The first person. In: G.E.M. Anscombe. The collected work. Vol 2.: Metaphysics and the philosophy of mind, 21-36. Oxford: OUP.

    Forchheimer, P. 1953. The category of person in Language. Berlin: de Gruyter.

    Goffman, e. 1979. Footing. Semiotica 25,1-2:1-29.

    Kantor, J. R. 1952. An Objective Psychology of Grammar. Bloomington/Evanston: Principia.

    Levinson, S. 1988. Putting Linguistics on a Proper Footing: Explorations in Goffman's Concepts of Participantion. In: P. Drew and A. Wootton (eds.). Erving Goffman: Exploring the Interaction Order, 161-227. Oxord Polity Press.

    L�vi-Strauss, Cl. 1962. La pens�e sauvage. Paris: Librairie Plon.

    Liebert, G. 1957. Die indoeurop�ischen Personalpronomina und die Laryngaltheorie. Lund: Gleerup.

    Majtinskaja, K. E. 1969. Mestoimenija v jazykx raznyx sistem. Moskva: Nauka.

    Majtinskaya, K. E. 1968. K tipologii genetic^eskoj svajzi lic^nyx i ukazatel'nyx mestoimenij v jazykax raznyx sistem. VJa 1968,3:31-40.

    Mauss, M. 1939. Une cat�gorie de l'esprit humain: La notion de personne, celle de 'moi'. JRAI 68:263-281.

    Mead, G. H. 1934. Mind, self, and society. Chicago: CUP.

    Myrkin, V.Ja 1964. Tipologija lic^nogo mestoimenija i voprosy rekonstrucii ego v indoevropejskom aspekte. VJa 1964,5:78-86.

    Russel, B. 1940. Inquiry into meaning and truth. New York: Norton.

    Schmidt, G. 1978. Stammbildung und Flexion der indogermanischen Personalpronomina. Wiesbaden: Harrassowitz.

    Schmidt, K. H. 1994. Zum Personalpronomen und der Kategorie 'Person' im Kartvelischen und Indogermanischen. HS 107:179-193.

    Schulze, W. 1998. Person, Klasse, Kongruenz. Vol. 1: Die Grundlagen. Munich: Lincom.

    Schulze, W. 1999. The diachrony of personal pronouns in East Caucasian. In: H. van den Berg (ed.). Studies in Caucasian Linguistics, 95-111. Leiden: CNWS.

    Schulze, W. 2003. The diachrony of demonstrative pronouns in East Caucasian. In: D. A. Holisky and K. Tuite (eds.). Current trends in Caucasian, East European and Inner Asian linguistics, 291-348. Amsterdam/Philadelphia: Benjamins

    Schulze, W. (forthcoming). Personhood in East Caucasian.

    ABOUT THE REVIEWER

    Wolfgang Schulze is the Head of the Institute for General Linguistics and Language Typology at the University of Munich. His main research topics include Language Typology, Cognitive Typology, Historical Linguistics, language contact, the languages of the (Eastern) Caucasus and Inner Asia, and 'Oriental' languages. He currently works on a Functional Grammar of Udi and on a comprehensive presentation of the framework of a 'Grammar of Scenes and Scenarios' in terms of 'Cognitive Typology'.