UTDtCSH.l 63.6'28; Vera Smole
Faculty of Arts, Ljubljana
A GEOLINGUISTIC EXAMINATION OF GENDER IN SINGULAR: NEUTER NOUNS IN -O IN SLOVENIAN DIALECTS
Using the methodology of linguistic geography (i.e., linguistic phonetic/morphological maps), and with the aid of dialectological literature and the material collected for the Slovenian Linguistic Atlas (Slovenski lingvistični atlas - SLA), this paper presents *-o reflexes and the geographical spread of the masculinisation and feminisation of neuter in Slovenian dialects, with an emphasis on the impact of phonetic changes on morphological changes. The paper draws a comparison with Ramov{'s earlier synthesis of these phenomena, and provides a more comprehensive explanation of feminisation in several northeastern Slovenian dialects.
S pomo~jo gradiva za nastajajo~i Slovenski lingvisti~ni atlas (SLA) in dialektolo{ke literature so z metodami lingvisti~ne geografije, tj. na jezikovni foneti~no-morfolo{ki karti, prikazani refleksi *-o in prostorski obsegi maskulinizacije in feminizacije nevter v slovenskih nare~jih s poudarkom na vplivu glasovnih sprememb na oblikoslovne. Narejena je primerjava s starej{o Ramov{evo sintezo teh pojavov in dodana nekoliko dopolnjena razlaga feminizacije v nekaterih slovenskih severovzhodnih nare~jih.
Key words: Slovenian dialects, neuter, word-final *-o, linguistic geography/geolinguistics, phonetic map, morphological map, masculinisation of neuter, feminisation of neuter
Ključne besede: slovenska nare~ja, srednji spol, izglasni *-o, lingvisti~na geografija oz. geolingvistika, foneti~na karta, morfolo{ka karta, maskulinizacija nevter, feminizacija nevter
1 When examining varieties of speech and attempting to detect the structural changes within them, it is important to determine the centres and the range of linguistic »innovation«; this allows one to recognise and understand the agents of those changes with greater ease. Without the knowledge of these changes and their range in dialects it is not possible to reconstruct the history of the Slovenian language as a whole or its standard variant. By focusing exclusively on written sources, only an incomplete and vague picture can be assembled (Bernstejn 22000: 302-306). Dialects are still the »basic source for the reconstruction of the development of the Slovenian language [because] they are, as a whole or in their individual elements, preserved in vastly different stages of development« (Logar 21996: 337-338), and are therefore a living reflection of history at any given moment. Linguistic geography, which began to develop at the end of the 19'h century in Germanic and Romance (specifically, French) philology,1 has enabled researchers to take a broad view of the history of language and it remains an ongoing and important development in dialectology. Unfortunately, linguistics has
1 Lucien Tesniere was the pioneer of Slovene linguistic geography and his Atl^s dvojinskih oblik v slovenščini (Atlas linguistique pour servir ä l'etude du duel en slovene, 1925) was the first to be drawn up for any Slavonic language. At the first International Congress of Slavonic Scholars in Prague in 1929, he was the first to suggest a compilation of a Slavonic linguistic atlas (along with A. Meillet). Two years later Tesniere set up the European Linguistic Atlas Organising Committee.
not paid enough attention to the results obtained by linguistic geography methodology (Bemstejn 22000: 300). The basic objective of linguistic geography is the compilation of linguistic maps for individual linguistic phenomena; these are then brought together in linguistic atlases that vary in terms of content and geographical range.2 In addition to grammars, manuals of official usage, and dictionaries, a national linguistic atlas is fundamental to any language. It is well known that the Slovenian language, in common with the other languages of the former Yugoslavia, is still without a linguistic atlas of its own, although the plan has now been seven decades in the making.3
2 Although contemporary Slovenian dialectology has the same research objectives as other linguistic disciplines - naturally, in close connection with them - 4 it must, at least in this writer's opinion, persist in its endeavours to complete the national linguistic atlas (working title - Slovenski lingvistični atlas / 'Slovenian Linguistic Atlas', or SLA), since it is an indispensable basis for other fundamental research work on linguistic history5 and dialectology. Efforts to complete the atlas are proceeding at the Dialectology Section of the Fran Ramovš Institute of the Slovenian Language, which is part of the Scientific Research Centre of the Slovene Academy of Arts and Sciences
2	Atlases can address one or more linguistic levels, or even a single linguistic category, can cover a single specific (dialect) area, the area of a single language, a country, several related or unrelated languages, etc.
3	The plan for the Slovenian Linguistic Atlas (SLA) was first outlined by Fran Ramovš in 1934. Francka Benedik has published, in the form of an introduction, a more detailed description of the progress of work on the atlas, the questionnaire (Ramovš's, later revised, supplemented, and alphabeticised), the network of research locations (Ramovš's, with later additions), additional field records made outside the network, field researchers, and a description of how the material and commentaries are organised ((Vodnik po zbirki narečnega gradiva za Slovenski lingvistični atlas (SLA)). This guide also contains an introduction by Karmen Kenda-Jež (pp. 5-9) outlining the origins of the SLA and the results of the work produced so far, with a focus on the points of departure of the research and the method used to collect and process the material. Given the volume of work - 406 research points and around 2.500 phonetic, lexical (a few semantic), and morphological questions - Ramovš's plan and subsequent minor amendments will be retained for the publication of the first lexical volume, planned for 2008, while the cartographical and commentary methodology will be consistent with current developments in (Slavonic) linguistic geography.
4	If the work on atlases is almost inevitably a collective endeavour (the only exceptions being atlases for specific areas), more recent work has also been done by individuals. Alongside a number of short dia-lectological works, this has also been demonstrated by several recent doctoral dissertations, e.g., Karmen Kenda-Jež, 2002: Cerkljansko narečje. Teoretični model dialektolo{kega raziskovanja na zgledu besedi{ča in glasoslovja. Doctoral dissertation supervised by Tine Logar, Ljubljana, Faculty of Arts. Computer print, 203 pp. + Appendices (Slovar osrednjecerkljanskega narečja (A-K), 156 pp.); Melita Zemljak, 2002: Trajanje glasov {tajerskega zabukov{kega govora. Instrumentalno-slu{na analiza. Doctoral dissertation supervised by Vera Smole and Zdravko Kačič, Ljubljana, Faculty of Arts. Computer print, 536 pp. + Appendices (published as Melita Zemljak, 2004: Trajanje glasov {tajerskega zabukov{kega govora: instrumentalno-slu{na analiza. Maribor: Slavistično društvo (Založba Zora; 30), 318 pp.; Danila Zuljan, 2005: Govorjena bri{ka nare~na besedila z vidika besedilne skladnje. Doctoral dissertation, supervised by Vera Smole and Simona Kranjc. Ljubljana, Faculty of Arts. Computer print, [236 pp.]
5	This is also evident in M. Greenberg '20 00, 220 02, whose Zgodovinsko glasoslovje slovenskega jezika needs numerous additions and amendments. These could be accomplished with a more consistent use of existing literature on Slovenian dialects and, more crucially, of the material collected for the SLA. One example of this is Greenberg's examination of short vowels in Chapter 40 ('Osredinjenje in izginotje visokih kratkih samoglasnikov (»moderna vokalna redukcija«)'/'Centring and disappearance of high short vowels (»modern vowel reduction«)') (2002: 161-165), the very title of which is misleading, as all vowels can become reduced and, later on, in certain positions disappear. Similarly, some short vowel phenomena in
(ZRC SAZU)6 and which houses all the material compiled so far. Over the years, many students of Slovenian language and linguistics at the Ljubljana Faculty of Arts have assisted in the collecting of material (Kenda-Jež 1999: 8).
3 To outline the methodology used in linguistic geography and the material collected for the SLA, the article will present - based on a few examples of neuter nouns in -o- the extent and the types of changes in Slovenian dialects in this marked, less stable morphological category of the Slovenian language, i.e only in the singular. In dialects this is such a complex problem that the existing material does not provide a complete explanation. It is thus necessary, at least in the more »problematic« areas, to record all originally neuter nouns and to determine the gender of each individual noun; this could only be accomplished in more detailed monographic descriptions of the morphology of local dialects. The spread of gender changes, i.e the transition from neuter to feminine (feminisation) or masculine (masculinisation), is also dependent on number. The plural differs from the singular in this regard, and both differ from the dual, since the very existence of the dual in some places is dependent on a previous change of gender (it is better preserved in the case of masculinisation), while the dual occasionally takes a singular and occasionally a plural base. The development of the neuter in Slovenian dialects is a very complex issue and one that merits special monograph treatment.7 This paper will deal with it only in part, primarily with the purpose of presenting the Slovenian Linguistic Atlas project and a small selection of the results obtained with the methodology of linguistic geography.
3.1 Unlike the masculine and feminine, the neuter is simply a grammatical gender. Moreover, in singular nouns the ways in which the neuter is expressed - in underived nouns primarily with endings, in tonemic variant also with accent - differ from those characteristic of masculine only in two cases with the same ending, i.e the nominative and the accusative. Many neuter nouns end in an unstressed -o (< *-o),8 which is one
Slovene dialects (some only concern unstressed vowels) not mentioned by Greenberg are important, e.g., akanje (*o and *e > a); ukanje (*o and *q > u - this is mentioned, erroneously, as being established in the western rather than the central dialects); positional ukanje and akanje (i.e., dependent on the consonant that comes before it); ikanje (*e and > i); e-akanje (*e and > a); umlaut after (functionally) soft consonants and before /j/ (*a, *o, *q > e); and even the diphthongisation of the word-final *-e, > -ie or *-o, *-q > -yo). These widely divergent dialectal developments in short vocalism are among the most characteristic innovations in Slovene and should have been more thoroughly discussed in Greenberg's work, which could have been accomplished with a closer examination of the existing dialectogical literature.
6	Associates of the Dialectology Section (Benedik, Jakop, Smole, Skofic, Poklac) have already published a large number of individual lexical/word-formation, morphological, phonetic, and accentological maps, as well as monographs (e.g., Jakop 2004; Poklac 2001). There has been recent strong collaboration with specialists in GIS (Geographical Information System) and electronic databases with the aim of preparing IT support for map-making, phonetic transcription, and the formulation of dialect dictionaries.
7	This would also be interesting from a comparative Slavonic point of view since the change in gender of various nouns through history is also attested in the dialects (and therefore the literary languages) of other Slavonic languages. The reasons for this can, however, differ widely.
8	The original *-o, i.e., the etymological -o, needs to be emphasised for the Slovenian language because in the central dialects (originally dialects of Dolenjska, Gorenjska, South Štajerska, and Eastern Rovte), the word-final *-o and *-p have developed differently; this con-/divergence will be shown elsewhere.
of the vowels in the word-final position that was first affected by modem vowel reduction. The interaction of phonetic levelling in the nominative/accusative singular and the large number of common endings in other masculine noun forms has in innovative dialects (and local dialects) caused the masculinisation of the neuter (the causes of feminisation need to be sought elsewhere, see 5.3). In geographical terms, both processes are today at different stages of development.
4 To show the (non)existence of the neuter across a range of contemporary Slovenian dialects9, the following nouns included in the questionnaire were taken from SLA material10 and their presence plotted on maps: 145" okno, 169 korito, 191 zito, 206 zelezo, 410 vino, and 517 /eto.12 The material was supplemented from other sources as necessity and opportunity dictated. The phonetic/morphological map Reflexes of the Word-Final *-o and the Gender of the Nouns 'okno', 'korito', 'žito', 'železo', 'vino', 'leto' shows the synchronic state of the -o ending (areas retaining the word-final -o and the reflexes arising from modern vowel reduction or morphological analogy) and the gender of these nouns in the singular. Phonetic changes in the word-final *-o are marked by isophones (see the map key),13 with the area attesting a preserved (or merely narrowed) -o (< *-o) lying outside the isophones. The gender of nouns is shown by means of hatching: slanted grey lines = an area that has preserved the neuter in the featured nouns; vertical black lines = masculinisation; horizontal black lines = feminisation; crosshatched lines = both phenomena).14 Map commentary is provided in section 4.2.
9	The criteria used for selecting these nouns were: (a) the existence of a lexeme across the whole area; (b) the non-positional nature of the development of -o (owing to its position after the consonant group, the noun okno is a partial exception here); (c) the usage of the word in everyday communication; (d) the unstressed nature of -o.
10	The material is too extensive to be dealt with here and can be consulted at the Dialectology Section of the Fran Ramov{ Institute of the Slovenian Language, ZRC SAZU, in Ljubljana. Material for 45 research locations in the SLA network (one for each dialect) can also be found in Poklac (2001: 26-36).
" The number before the noun signifies the serial number of the question in the SLA questionnaire (in: Benedik 1997: 26-86).
12	As will become clear later, the ending itself does not always indicates the gender of the noun. In Slovene, neuter nouns can, following the umlaut, also end in -e, which is why we verified the gender using two further questions from the SLA questionnaire (813 masculinisation of the neuter and 814 feminisation of the neuter) which required adjective + noun (veliko okno, močno sonce) or noun + verb in the past tense (okno je bilo zaprto, sonce je sijalo) types. Unfortunately, the material relating to these two questions is also very deficient.
13	If a line bisects the number of the research location, this signifies that both (or even three) *-o reflexes in contact are present in the local dialect, e.g., at point 256 -o can be preserved or reduced. In the first case the noun is neuter, in the second case masculine; given this, the local dialect is transitional. # signifies the zero reflex (*-o > -0).
14	Only those local dialects in which changes have occurred in one of the featured nouns or in other neuter nouns have been excluded. Occasionally the gender of a noun changes independently of the general tendencies of neuter change; one of the most frequently used is jabolko 'apple', where this noun is also feminine in Southern Štajerska local dialects with masculinisation in the singular, i.e., jabka, which could be by analogy with hruška 'pear' and with other common types of fruit, which are all feminine (češnja, češplja, sliva, marelica, and the re-formed breskva, etc.).
4.1	That developments in short unstressed vowels are relatively recent, particularly in non-central dialects, is confirmed by the fact that the isoglosses of individual developments do not correspond with the geographical boundaries of dialects. At the same time, the map neatly shows the expansion of phonetic innovations from the centre towards the periphery, as well as several different focal points of morphological innovation. Of course, a single map, though it might present the range of phenomena,15 can by no means show the full complexity of the changes undergone by the neuter, even in the limited number of nouns. Even the SLA material is somewhat deficient as regards morphological questions; a proper understanding of the morphology of Slovenian dialects (and therefore of accent) will require more research detailing the features of individual local dialects, as has been pointed out on a number of occasions (e.g., Orožen 20032: 184). The findings outlined below are merely an attempt to initiate a new synthetic analysis of the neuter in Slovenian dialects.
4.2	As regards the level of preservation of the neuter singular in the nouns okno ('window'), korito ('trough'), žito ('corn'), železo (' iron'), vino ('wine') and leto ('year'), which are a fairly satisfactory representation of neuter nouns in -o, Slovenian dialects can be divided into four groups: 1) the neuter has been preserved; 2) the neuter has been masculinised; 3) the neuter has been feminised; 4) according to a set of rules (or none) some nouns have remained neuter while others have been feminised or (more rarely) masculinised. In a very rare number of cases, both feminisation and masculinisation are possible (see 5.2). Since gender change is strongly (though not absolutely) connected with the phonetic ending of the noun, the neuter will be presented in relation to what happens to -o:
4.2.1 The neuter is preserved if *-o:
4.2.1.1 is preserved (*-o = -o) or partly narrowed (*-o > -o, -o, -o), as in the Koroška Zilja and Western Rož local dialects in Austria,16 similarly in the Eastern Podjuna and Western Northern Pohorje and most of the Mežica dialects. In the Pannonian dialects, it is more consistently preserved in the local dialects of Prekmurje (with the exception of the southern local dialects), as well as in the Eastern Slovenske Gorice dialect, and in the Haloze, Prlekija and Štajerska Central Štajerska and Kozjansko-Bizeljsko on
15	The most common practice in modern linguistic atlases is to use a single map to show a single phenomenon in an individual word. In this case this would mean producing six phonetic and six morphological maps. Identical or related phenomena can be shown at the end in general maps.
16	The names of the dialects have been taken from the Karta slovenskih narečij (Map of Slovenian Dialects, Logar - Rigler 1983), with the exception of the Goričansko dialect, which is referred to as the Slovenske Gorice dialect (after Koletnik 2001). The dialect group to which the dialect or parts of the dialect (local dialects) belong is always given in italics, mostly preceding mention of the local dialects (but occasionally after, as style dictates). In this case, 'Koroška' denotes the dialect group to which the Zilja and Ro' local dialects belong; since in this case the phenomenon does not cover the whole dialect area, the local dialects are given for the purposes of greater precision, usually with its compass definition (north-east, south-west, etc.).
the border with Croatia.'7 In the west it is preserved in the Primorska Ter, Nadiža, and Brda local dialects, and in most of the Soča, Banjsice, Karst, Istria, and Čičarija local dialects. Within the Dolenjska dialect group, the Bela Krajina and Eastern Kostel local dialects preserve -o and the neuter;
4.2.1.2	is narrowed to -u (*-o > -u), as in the Primorska Notranjska dialect and in the Karst and Istrian local dialects in contact with it; in the southern part of the Dolenjska dialect (the Eastern Dolenjska local dialects from Novo Mesto towards Gorjanci); in the local dialects of most Štajerska dialects; in a more condensed fashion mainly in the Southern Pohorje and Central Savinja local dialects; and in the Panonnian dialect group, where the Southern Prekmurje local dialects have -u or (slightly reduced u);
4.2.1.3	is aspirated (*-o > -o), as in the Primorska Rezija local dialects (with the exception of the Bila local dialect);
4.2.1.4	is diphthongised (*-o > -uo),'8 as in several Primorska Karst (Lower Vipava Valley and along the coast) and Central Istria local dialects;
4.2.1.5	is partly reduced to -a (*-o > -a, probably through an intermediate -u), as in the Koroška Eastern Rož and Obir local dialects; in the Western Podjuna and Štajerska Upper Savinja local dialects'9; and in Dolenjska local dialects south of Ljubljana (in the latter, for the most part only positionally).20
4.2.2 Neuter nouns are masculinised in the singular if the -o is completely reduced (*-o > -0, through an intermediate -u, which can still be preserved positionally);2' this feature appears in all Gorenjska dialects, in the Rovte Črni Vrh, Poljanska Sora, Horjul, and Škofja Loka local dialects, in the Cerkno Lanise local dialect (168), in northern Dolenjska and most of the Eastern Dolenjska local dialects (Dolenjska)22, in most local dialects of the Central Savinja and Posavje dialects, in the Western Koz-
17	In all Pannonian and Štajerska local dialects listed at least some of the featured nouns are feminine (frequently okno for example, or the nouns more commonly used in the plural, since the plural is mostly feminised here).
18	This diphthong has undergone further development, i.e., -uo (103 Renče); -uo (104 Branik); -uä (112 Križ/S. Croce); -ua 113 (Prosek/Prosecco); -ua (118 Dekani, 119 Kubed). The numbers here refer to the number of the location in the network of research points.
19	One exception is the Mozirje dialect, where the post-war generation abandoned the old -a ending and replaced it with -o (compare the field record from 1952 with those from 1966 and 1973).
20	Consistent *-o > -u > -a development is found in local dialect 231 Rakitna (stie-.gna, vi-.na), and positionally, after the consonant group, in: 240 Grosuplje (stä-.gna : vi:n), 253 Gorenje Brezovo (ste:gns), 245 Stična (stä.-gna, sä:dls), 255 Muljava (steä-.gna). In the first case, the nouns are neuter, in the second case they are masculine. This is marked on the map with an asterisk to the right of the number of the location.
21	Marked on the map with an asterisk to the right of the number.
22	In local dialect 262 Šentrupert, for example (see Smole 1997: 170), the -u is preserved for the consonant group ending in l or v, while the noun nevertheless remains masculine (usak stja:blu 'visoko steblo', an dja:blu jepau despuot eno deblo je padlo 'čez pot', ta:k buga:st(v)u 'tako bogastvo').
jansko-Bizeljsko and Central Štajerska local dialects (Štajerska), and in the South Eastern Rož and Western Obir local dialects (Koroška).
4.2.3	Some neuter nouns are feminised in the singular23 if by analogy (Ramovš 1952: 37) they take an a-declension ending.24 This has happened at the point of contact between the Koroška, Štajerska and Pannonian dialect groups, i.e., to the greatest extent in the Kozjak and Southern Pohorje local dialects of the Štajerska dialect and in the Pannonian Slovenske Gorice dialect, to a lesser extent stretching further towards the nearby Prlekija local dialects on the one side and to the Northern Pohorje local dialects (Koroška) on the other.25 Z. Zorko (1995,26 1998) and M. Koletnik (2001: 129-132) have described the complex situation that has arisen in relation to these local dialects on a number of occasions. The other, less significant focal point of this phenomenon, caused by akanje, is the point of contact between the Primorska and Rovte dialect groups along the Sava river and its tributary, the Bača river, i.e., in two Soča local dialects (72 Zatolmin, 73 Ciginj), one Tolmin local dialect (161 Most na Soči) and in all Bača local dialects27 (158 Rut, 159 Podbrdo, 160 Porezen).
4.2.4	The exceptions, where the phonetic ending conceals the actual gender,
are as follows:
4.2.4.1 *-o > -a following akanje, neuter is preserved: in two local dialects of the Banjšice micro-dialect of the Primorska dialect group (90 Avče, 95 Lokve); in two Tolmin local dialects of the Rovte dialect (162 Grahovo ob Bači, 163 Čepovan); and in one Cerkno local dialect (166 Cerkno). Two local dialects close to Cerkno are, on account of their proximity, proper mixed local dialects: in the first (165 Šebrelje), several nouns are neuter despite their -a ending, while others are feminine (gender is distinguished by the genitive); in the second (164 Gorenja Trebuša), the same noun can be masculine and feminine (the two forms co-exist);
23	According to Ramovš (1952: 36-37), the focal point of this phenomenon should be Kozjak and Goričansko. The Prekmurje dialect is spoken in Goričansko, where feminisation is not attested except in a very small number of nouns. Generally speaking, Ramovš indicates a wider area of feminisation than that indicated by the SLA material. There are two possible reasons for this: either he did not have accurate data and reached his conclusion on the basis of individual words; or the area covered by this phenomenon has shrunk. This last possibility appears to be confirmed by local dialect 387 Cankova (Prekmurje). According to Pavel (1909), this local dialect had several feminised nouns (bedro, gnezdo, čudo), others were still neuter (delo, leto, mesto), and okno was both feminine and neuter. At the present time, all of the above nouns are neuter (fieldwork performed by T. Jakop, 12 Nov 2005).
24	The following nouns are most frequently feminised: koleno, korito, kopito, okno, rebro, stegno. Leto, vino, železo, žito, or nouns used primarily in the singular (leto is an exception here), retain the neuter and the (narrowed) -o ending. Individual nouns with an -a ending can still be found in some Prlekija local dialects, and more rarely in the local dialects of Southern Pohorje and Northern Pohorje. A similar situation as that found in Kozjak and Slovenske Gorice local dialects is encountered in the neighbouring Northern Pohorje local dialect of Koroška (55 Zgornje Kaple).
25	The only exception is local dialect 367 Negova (-p ending and a preserved neuter).
26	Our material does not always agree with hers.
27	This is the dialect of Slovenicised Germans, who have, generally speaking, greatly simplified the morphology.
4.2.4.2	*-o > -a following akanje, while the noun is masculine: in two local dialects of the Koroška Zilja dialect (5 Ukve/Ugovizza,28 8 Rateče);
4.2.4.3	*-o > -0, the noun preserves the neuter: in the Koroška Rož local dialect (28 Žitara vas - Sittersdorf);
4.2.4.4	*-o > -u or -a after a consonant group, but the noun is nevertheless masculine (see 1.2.2 and 1.2.1.5).
5 The most condensed description of the neuter in Slovenian can be found in Ramovs's Morfologija slovenskega jezika (Morphology of the Slovenian Language, 1952: 35-37). This work features a number of inaccuracies, and is often at variance with his findings in his Dialekti (1935), which is a further proof that his Morfologija was only an approximate indication of what he knew at the time.29
5.1 Ramovs noted all the phenomena described here; the discrepancies with our findings lie chiefly in their geographical extent and have arisen as a result of the expansion of the modern vowel reduction of the word-final -o in the last 50 years (it should also be borne in mind that this paper is restricted to only a few -o nouns). Ramovs does not make further mention, for example, of tendencies towards the narrowing of -o to -o, -o (and even to -u) in a number of Štajerska and Pannonian local dialects (which he believed still preserved -o). One also observes a widening of the area in which a complete -u (< *-o) > -0reduction is attested: to the Western Gorenjska local dialects, several Northern Dolenjska local dialects, and to most of the Eastern Dolenjska local dialects. Ramovs (1952: 36) was well aware that masculinisation was taking place for morphological reasons, since the forms differed only in the nominative and accusative, and that the complete reduction of the nominative and accusative ending and the equalisation of all singular case endings were even further accelerating the process. The situation we see today is also proof of this, since masculinisation is known in all local dialects in which the -o (along with a fair number of others) has disappeared and now covers a condensed area along the Sava river from the Kanalska Valley in Italy, across the whole of Gorenjska and through the Karavanke deep into Koroska (the Rož part), and to Obir, the whole of Sevska and Poljanska valleys, the Črni Vrh plateau, the Ljubljana basin, the northern half of Dolenjska (including Posavje and Zasavje) and Southern Štajerska along the Savinja to the mouth of the Paka, and the Celje basin up to Šmarje pri Jelšah. The morphological nature of masculinisation is also confirmed by its expansion to the -o or -a < *-o area to two Koroška Rož local dialects
28	E.g. truipwa, kalema, žaleiza.
29	Ramovs's Morfologija slovenskega jezika (Morphology of the Slovenian Language) is an edited collection of lecture notes from 1947/48 and 1948/49 prepared by his students B. Pogorelec, P. Merku, and M. Sovre. The author reviewed and approved the notes on 1 June 1952, when his health was already failing (he died three months later). Although this work remains invaluable, its very nature (a series of lecture notes) means that it does not match the quality of his monographs and does not reflect the full extent of his findings.
(16 Sveče [Suetschach] and 19 Slovenji Plajberk [Windisch Bleiberg]) and two Zilja local dialects (8 Rateč and 5 Ukev) with -a < *-o, while the Rož local dialect 28 Žitara vas (Sittersdorf) preserves the neuter despite a silent -o. Nouns with positional -u or -9 do not preserve the neuter after the consonant group in a number of Dolenjska and Eastern Dolenjska local dialects.30
5.2	Ramovš was not aware of feminisation of the neuter in the west, in the Bača micro-dialect of the Rovte Tolmin dialect,31 which through feminisation (Logar 21996: 412) in the plural, and as a result of akanje, has spread to neighbouring local dialects via individual nouns; it has already spread to the Tolmin (161 Most na Soči) and Cerkno local dialects (165 Sebrelje). Feminisation is also attested in the 164 Gorenja Trebuša local dialect, although masculinisation is also advancing there from the east such that both phenomena are possible. Ramovš was aware of local dialects which, despite akanje, preserved the neuter (e.g. other Cerkno and Karst Bača local dialects). In general, the area of contact between the Rovte and Primorska dialect groups is a very sensitive one as far as this phenomenon is concerned and the present situation probably also unstable. Phonetic developments do not overlap with morphological developments; rather, phonetic and morphological phenomena occur in contact with each other. Further changes can be expected.32
5.3	Ramovš (1952: 37) describes a feminised area in the east that is only slightly larger than the mixed area here, and he incorrectly gives Goričansko as the centre of this phenomenon, alongside Kozjak. It was clear to Ramovš in the Dialekti (1935: 172, 175, 190) that feminisation in the Pannonian Prekmurje dialect and the northeastern part of the Prlekija dialect was confined to a few neuter nouns, which could differ from one local dialect to the next. In none of the local dialects in this dialect does feminisation encompass all nouns; the preservation of a number of neuter nouns, which Ramovš attributes solely to Slovenske Gorice in the Morfologija, in fact applies to the whole area. The rules of feminisation of individual nouns in the Prekmurje and Prlekija dialects are difficult or indeed impossible to ascertain (it is partly the case that feminisation is more common in those nouns generally used in the plural). At the centre of this occurrence (in the north of Slovenske Gorice, i.e., the Slovenske Gorice dialect, as well as in the Kozjak and Pohorje local dialects), one can detect some of the rules by which the new morphological division of neuter nouns has arisen. According to Koletnik (2001: 129-132), these are as follows for the Slovenske Gorice dialect: some neuter nouns can preserve their gender and declension only in the singular, being feminised or masculinised in the dual and plural; in the singular the neuter is mostly preserved (1) in uncountable nouns (e.g., mleko, vino, žito), (2) in nouns which in oblique cases extend the stem with -t and -n; however, some of them, because they
30	Ramovš does not mention Dolenjska dialects in which -o > -u > -3 reduction is possible, although it is now attested in Rakitna and to the east of there; we can infer from this that it is a recent development.
31	Rigler (22001: 450) has drawn attention to this: »Feminisation of the neuter also appears in the west, not just in the east (see Cronia, SR III, 324).«
32	On the microlocations of individual phenomena, see Kenda-Jež (1999a: 217, footnote 12).
also have the extension in the nominative by analogy, can also be masculine (e.g., tele, zrebe; težko bremen and tisti bremen); (3) a considerable number of nouns take the -a ending in the nominative and decline according to the feminine -a declension, or are feminised in the singular (e.g., cela, gnezda, kolena, korita, sila, okna). In several Slovenske Gorice local dialects (Negova, Ivanjci, Črešnjevci, and Radenci), certain nouns that are neuter in the nominative are then declined as feminine nouns (koleno, -e, -i). In the Morfologija (1952: 37), feminisation in this area is explained as being the result of pluralia tantum (pluCa, jetra, nebesa, vrata), where, following akanje in the dative, locative and instrumental endings -om > -am, -ah, -ami, the majority of plural endings were the same as the feminine -a endings, as expected, which is why the nominative took -e (pljuce, jetre) by analogy, with a gradual transition to feminine declension in the plural and then the singular. The weakness of this explanation (accepted by Zorko and Koletnik) lies in the fact that akanje has not been attested in the local dialects in this area, as Ramovš well knew. He does not explain the generalisation of -a endings by akanje but by analogy with the nominative and accusative plural (Dialekti, 167).33 The commonly used genitive zero-ending had an influence here that should not be overlooked, as did the syllable word structure present in both feminine -a and neuter declensions (lipe, lip = vrata, vrat).^'4 The original difference in the nominative is not important since nouns of the same gender can also have different phonetic endings in the singular (already in neuter: in -o or -e).
6 Conclusion. Using the methodology of linguistic geography to show the extent of preservation of neuter nouns in -o in Slovenian dialects, and by comparing these findings with those of Ramovš from an earlier date, the following can be said:
1)	The process of masculinisation in central Slovenian dialects (in a wide belt that stretches along the Sava river) is ongoing, and is closely connected with a phonetic phenomenon, i.e., reduction *-o > -0;
2)	The process of feminisation has two centres: (a) the north-east, in the wider surroundings of Maribor, which has its origins in morphological analogies in the plural and has affected only some neuter nouns; with lexicalisation, the process stopped not long ago, and the situation can differ widely from one local dialect to the next; (b) the west, in the western environs of Tolmin, where it is the result of akanje and where the process is still continuing and is also being stimulated by the tendency towards simpli-
33	The following passage from the Dialekti (p. 167) confirms that Ramovš's summary was wrong: »Of the morphological features [Pohorje-Kozjak dialect, author's note], particular mention should be made of the feminisation of the neuter, which to a greater or lesser extent appears in Kozjak, in Slovenske Gorice, and in Prekmurje; it originates from the unification of the plural in the feminine and neuter: following the nominative and accusative -a, -am, -ax, -ami also appeared in the neuter (the genitive was without an ending); and now, following the feminine (which has had the same form, except in the nominative and accusative), a form has arisen with -e in the nominative and accusative: vräte (and all the neuter plural only nouns) are now perceived as being feminine singular; e.g., ucesa for oko, ena drevesa, ena uäkna (leto, drevo, mleigku, üxo, etc. are also used).«
34	In the Eastern Dolenjska local dialect of Sentrupert (262), following masculinisation of neuter, the stem of the genitive plural began to extend with -ov (korakov = koritov).
fication of declensions; the situation differs from one local dialect to the next, and the vowel ending in the nominative singular is not relevant to the gender of the noun;
3) The neuter has been preserved in a fairly wide belt on the periphery; the vowel ending has generally been preserved, albeit in a partly reduced form.
V angleščino prevedel Joel Smith.
References
Benedik, f., 1999: Vodnik po zbirki nare~nega gradiva za Slovenski lingvisti~ni atlas (SLA). Ljubljana: Založba ZRC (ZRC SAZU), 154 p.
Bernštejn, s. B., 22000: Razmyšlenija o slavjanskoj dialektologii. In: Iz problematiki dialek-tologii i lingvogeografii. Sbornik statej. Moskva: »Indrik«. 299-309.
Jakop, t., 2004: Dvojina v slovenskih nare~jih. Doktorsko delo, mentorica Vera Smole, somen-tor Vladimir Nartnik, Ljubljana, FF. Computer printout, 213 p. + Priloge (indeksi: 55 p., member of maps: 26).
Kenda-Jež, k., 1999: Predgovor. In: F. Benedik, Vodnik po zbirki nare~nega gradiva za Slovenski lingvisti~ni atlas (SLA). Ljubljana: Založba ZRC (ZRC SAZU), 5-9.
—	1999a: Fonološki opis govora kraja Cerkno(OLA 6, SLA 166). Jezikoslovni zapiski 5: Zbornik Inštituta za slovenski jezik Frana Ramovša. Ljubljana: Znanstvenoraziskovalni center SAZU, 207-234.
Koletnik, m., 2001: Slovenskogoriško nare~je. Maribor: Slavistično društvo (Zora; 12), 323 p.
Logar, t., 21999: Rovtarska narečja (Članki za Enciklopedijo Slovenije). In: Dialektolo^ke in jezikovnozgodovinske razprave (ur. Karmen Kenda-Jež). Ljubljana: ZRC SAZU, Inštitut za slovenski jezik Frana Ramovša, 407-412.
Orožen, M. 22003: Pogled na slovensko narečno oblikoslovje od V. Oblaka do T. Logarja. In: Razvoj slovenske jezikoslovne misli. Maribor: Slavistično društvo (Zora; 26), 173-184.
POKLAČ, S., 2001: Akanje v slovenskih nare~jih (po gradivu za Slovenski lingvisti~ni atlas). Diplomsko delo, mentorica Vera Smole, Ljubljana, FF. Computer printout, 119 p. + 12 maps.
Ramovš, F., 1935: Histori~na gramatika slovenskega jezika VII., Dialekti. Ljubljana: Učiteljska tiskarna, 204 p.
—	1952: Morfologija slovenskega jezika. Skripta, prirejena po predavanjih prof. dr. Fr. Ramovša v l. 1947/48, 48/49. Ljubljana: DZS, 167 p.
Rigler, j., 22001: Ramovševa Morfologija slovenskega jezika. In: Zbrani spisi 1. Jezikovnozgodovinske in dialektološke razprave. (ed. V. Smole). Ljubljana: Založba ZRC, ZRC SAZU. 449-464. [Written in 1953/54, published in SR 33/3. 1986.]
Smole, V., 1997: Sovplivanje samoglasnikov in soglasnikov v vzhodnodoelnjskih govorih. Jezikoslovni zapiski 3, Ljubljana: Znanstvenoraziskovalni center SAZU. 167-173.
ZORKO, Z., 1995: Nare~napodoba Dravske doline. Maribor: Kulturni forum (Zbirka Piramida; 3). 362 p.
—	1998: Haloško nare~je in druge dialektološke študije. Maribor: Slavistično društvo (Zora; 6). 342 p.
Povzetek
Srednji spol je v sodobnih slovenskih narečjih tista kategorija, ki izkazuje različno stanje: ohranjenost, maskulinizacijo ali feminizacijo, obseg teh pojavov pa ni natančno določen ne v oblikoslovnih sistemih ne glede na prostor. Droben prispevek k temu je z metodami lingvistične geografije obdelan spol samostalnikov srednjega spola na nenaglašeni -o (okno, korito, žito, Železo, vino in leto) - gradivo je iz zbirke za Slovenski lingvistični atlas (SLA), ugotovitve pa primerjane s sintetičnimi starejšimi (Ramovš 1952, 1935), ki zajemajo vse samostalnike srednjega spola. Prostorski prikaz in razlage pojavov maskulinizacije in feminizacije nevter so v Ramovš 1952 netočne, mestoma napačne, a žal v dialektološki literaturi najpogosteje citirane; pravilneje, a seveda v delu razpršeno in veljavne za tisti čas, so pojavi prikazani v Ramovš 1935. Ugotovitve, nastale s primerjavo slednjih in dela kartografiranega gradiva za SLA, lahko strnemo takole:
1)	proces maskulinizacije v osrednjih slovenskih narečjih v širokem pasu ob Savi se nadaljuje in je tesno povezan z glasoslovnim pojavom, tj. z redukcijo *-o > -0;
2)	proces feminizacije ima dve središči: a) severovzhodno, v širši okolici Maribora, ki ima vzvod v morfoloških analogijah v množini in je zajel samo del samostalnikov srednjega spola - z leksikalizacijo se je proces nedolgo nazaj ustavil, stanje v posameznih govorih pa je lahko zelo različno; b) zahodno, v zahodni okolici Tolmina, je posledica akanja, proces še poteka, spodbuja pa ga tudi težnja k poenostavitvi sklanjatev - stanje v posameznih govorih je različno, glasovna končnica v I. ed. za spol samostalnika ni relevantna;
3)	srednji spol se ohranja v dokaj širokem obrobnem pasu; glasovna končnica, četudi v delno reducirani obliki, je praviloma ohranjena.