1The maritime vocabulary of Croatian coastal and insular vernaculars has received remarkable scholarly attention due to its composite nature, which reflects a centuries‑long process of Mediterranean acculturation of the Adriatic Slavs. This process included extensive lexical borrowing from Greek, autochthonous Romance varieties (the so‑called “Dalmatian Romance”), and various Italo-Romance dialects, foremost Venetian. This is particularly true in the lexicon of marine fauna, with Vojmir Vinja’s remarkable etymological and motivational study (JaFa) serving as the reference point for all subsequent research. While Vinja’s extensive fieldwork in the 1950s admittedly owes some debt to the classical linguistic geography, his work primarily focused on the structural-motivational aspects of dialectal vocabulary rather than on the geolinguistic interpretation of its spatial distribution (cf. Bazina, 2024: 34–37). Drawing from Vinja’s experience and Deanović’s fieldwork for ALM (e.g. Deanović, 1959, 1966, 1967), the past two decades have seen several research initiatives more in line with the geolinguistic methodology. One of these, the Linguistic Atlas of Maritime Culture in Dalmatia and Kvarner (JAPRK), was first presented in this journal (Vuletić & Skračić, 2016). By the completion of the fieldwork, the JAPRK survey network has been significantly expanded to meet the growing interest of local communities (Vuletić et al., 2022: 200–202). In the semantic field of marine fauna, where greater diatopic variation was expected than in other sectors of the maritime vocabulary, JAPRK has been supplemented with microareal research based on an exceptionally dense network (cf. here 2). With the first series of JAPRK onomasiological maps still in preparation, this paper deals with several semasiological aspects of Croatian vernacular vocabulary of marine fauna, much in the spirit of the qualitative geolinguistic analysis proposed by such scholars as Jaberg (1936), Terracini (1963, 1964), and Telmon (1989). In doing so, we also address an important methodological issue concerning network density, i.e. the complementary relationship between macroareal and microareal tools in linguistic geography.
- 1 Jaberg (1936: 47) speaks of types lexicologiques. One might refer to such lexical types as vocables(...)
- 2 From the diasystemic perspective (cf. n. 1), one could also encode them as types ⸢ugor 1⸣ ‘Anguilla (...)
2The five semasiological maps examined in this paper present five lexical types, recorded in Croatian vernaculars of Dalmatia and the Kvarner Gulf, that have undergone semantic differentiation, with different meanings covering different geographic areas, or, as Jaberg (1936: 42–77) called them, aires sémantiques. As such, these types are not the result of linguistic analyses but rather forms referring to the object of the analyses (cf. Bazina, 2024: 23, no. 18), in this case a lexical meaning.1 We have chosen predominant vernacular realisations as type forms, an option frequently used in the process of lexical typification (cf. Regis, 2019). Each of our semasiological maps features geographically differentiated meanings of a given lexical type, represented by different colours and rendered by the respective taxa, e.g. the type ⸢ugor⸣ displays the meanings ‘Anguilla anguilla’ and ‘Conger conger’.2 These maps are based on combined data from the respective onomasiological maps, which reflect the concepts of the JAPRK questionary (Q)—in this case animal species. For the sake of transparency of our analysis, the semasiological maps are accompanied by the respective onomasiological maps with typified data, following the principle of the predominant vernacular realisation, except in the case of ⸢pizdica⸣ (see further below).
- 3 The concept of cultural vocabulary is notoriously hard to define, as it is language specific. For a (...)
3Following the guidelines set out by Jaberg (1936: 44), the general aim of our maps analysis is to shed light on the conditions of formation, concurrence and geographical distribution of different meanings (cf. Jaberg, 1936: 44). With maritime lexicon representing an instance par excellence of what has been labelled cultural vocabulary, lexical areas are interpreted as the result of choices made by different points within a common cultural and social setting (cf. Terracini, 1964: 167).3 Such an approach underlines the eminently historical character of linguistic geography, as pointed out by several authors, most eloquently by Telmon (1989).
4Although Jaberg (1936: 42) stated that the linguistic geography had “no pretension to create a corpus of particular doctrines”, he nevertheless endeavoured to formulate some principles of the interpretation of linguistic maps—a kind of methodological backbone of what he called “appreciation of linguistic probabilities”. We are, for our part, fully convinced of their heuristic value. In fact, as we shall see, four of our maps correspond simultaneously to Jaberg’s typology of “aires exclusives” and “aires contiguës” (or almost), while one, that of ⸢pizdica⸣, fits the case of “types lexicologiques décadents”, that is, of “aires sémantiques disséminées d’un type lexicologique en recul” (cf. Jaberg, 1936: 44, 45, 47). That being said, Jaberg’s masterly analysis is based on the materials of ALF and AIS, classical examples of macroareal (= großräumige, cf. Jaberg, 1955) linguistic atlases. Findings relying on such geolinguistic materials may risk overgeneralization due to their relative network density. As Camproux (1962: 760) observed, “on s’est satisfait d’une vue d’aigle”, while “l’humble attitude du ver luisant qui éclaire sa lanterne à hauteur de ventre” would have enabled us to observe the real course of dialectal boundaries.
- 4 The geolinguistic structure is to be understood as a spatial relation between network points in ref (...)
5A denser survey network “may help to minimize validity and reliability problems” (König, 2010: 499), inherent to the geolinguistic method. It can also complicate observations on clear‑cut limits of “semantic areas”. Jaberg was fully aware of this problem, as demonstrated in his essay on “Großräumige und kleinräumige Sprachatlanten” (Jaberg, 1955), yet he tended to identify macroareal atlases with national atlases. While JAPRK is undeniably a regional linguistic atlas, it nevertheless covers a considerable space, stretching over roughly 90% of Croatia’s 5,835.3 km‑long coastline (more than the total length of the mainland coast of France, as a point of comparison). Previous experience, notably Vinja’s fieldwork for JaFa, suggested that the JAPRK’s survey network, designed to capture the essential geolinguistic structures of the maritime lexicon in general, might fail to do so in the case of marine fauna.4 To achieve a more informed geolinguistic analysis in this sector, supplementary fine‑grained fieldwork, almost “point par point” as advised by Camproux (loc. cit.), was conducted with the same questionnaire in Northern Dalmatia (Vuletić & Skračić, 2018) and Central Dalmatia (Bazina, 2024) and is now underway in the Kvarner Gulf (Vuletić & Bazina ms.). The maps presented in this paper are thus based on multiple, yet methodologically homogenous sources. To distinguish these sources, JAPRK points are marked by circles, and those from our supplementary research by squares. In addition, some of Vinja’s points from JaFa, marked by triangles, were included to provisionally complete our ongoing research in certain areas. The following analysis concerns not only the geolinguistic interpretation of the maps but also the methodological interface of macroareal and microareal research, which together constitute the combined basis of our maps. We refer to the geolinguistic structures reflecting the basic JAPRK network as macrostructures, and to those reflecting the supplementary network as microstructures.
6Note: Throughout this paper, we will refer to the traditional terms Čakavian and Štokavian, denoting two macrodialects present on the Croatian coast, although they admittedly lack solid grounding in terms of historical and comparative linguistics. The same applies to the label Dalmatian Romance, used broadly to denote autochthonous Romance varieties of the Dalmatian coast, from the Kvarner Gulf to the Bay of Kotor, characterized by conservative features. We will make frequent use of the following abbreviations: Cr. – Croatian, DalmR. – Dalmatian Romance, ER. – Early Romance, Gr. – Greek, It. – Italian, Lat. – Latin, Ven. – Venetian, PSl. – Proto‑Slavic.
Figure 1. – Map S‑1. ⸢ugor⸣.
- 5 Some sources (e.g. Hirtz, 1956: 440) document ⸢ugor⸣ ‘eel’ in Herzegovina, but no geolinguistic dat (...)
7Map S‑1. ⸢ugor⸣. Based on maps O‑1. Anguilla anguilla (Q383) and O‑2. Conger conger (Q385) (cf. appendix). A rare case of Slavic fish name transplanted from the continental Slavic Urheimat to the Adriatic shores, ⸢ugor⸣ < PSl. *ǫgorь (ERHJ II, s.v. ȕgor; cf. ERHSJ III, s.v. ugor; JaFa I: 266) has experienced a threefold fate: it has retained the original Slavic meaning ‘eel’ on the coast and the inner islands (Krk, Rab, Pag) of the Northern Adriatic; it has come to designate another eel‑shaped fish, the European conger, in most of the Central Adriatic; it has disappeared in Southern Dalmatia, as well as on the outer Kvarner islands and in the Zadar archipelago, except for its southernmost fringes.5 This is due to the presence of several Romance loanwords. For Anguilla anguilla (map O‑1): ⸢angulja⸣ and ⸢jegulja⸣, both ultimately from Lat. anguīlla, the latter quite old, judging from the denasalisation an# > ę > e and its presence elsewhere in the South Slavic area (ERHSJ I, s.v. jegulja; ERHJ I, s.v. jegulja; Ligorio, 2014: 77), the former somewhat younger, with no denasalisation and exclusively Adriatic; ⸢bižat⸣ < Ven. bisoto (ERHSJ I, s.v. bižot; Ja Fa I: 157), originally a designation for smaller eel (Vuletić & Skračić, 2018: 41–42).
Figure 2. – Anguilla anguilla (Map O‑1).
8For Conger conger (map O‑2): ⸢grug⸣, ⸢gruj⸣, ⸢grum⸣, ⸢grun⸣, and ⸢grunj⸣, all presenting areally specific developments of the DalmR. reflexes of Lat. conger, or rather ER. *congru (ERHSJ I, s.v. grug; Ligorio, 2014: 116).
Figure 3. – Conger conger (Map O‑2).
- 6 Štokavian vernaculars of continental origin, introduced to the shores of North Adriatic during the (...)
9The reason for the loss of ⸢ugor⸣ in much of the (continental) South Slavic area must have been the early borrowing of anguīlla, while the necessity to distinguish between eel and conger on the Adriatic shores likely contributed to its preservation on the western periphery. This preservation followed different solutions: in the Northern Adriatic, the retention of ⸢ugor⸣ ‘eel’ and the borrowing of Romance designations for conger; in the Central Adriatic, the semantic shift of ⸢ugor⸣ to ‘conger’ and the borrowing of Romance designations for eel. In this way, the Adriatic area of ⸢ugor⸣ is made of two exclusive semantic areas, in Jaberg’s terms, that of ⸢ugor⸣ ‘eel’ and ⸢ugor⸣ ‘conger’. These areas are geographically separated yet very close to one another. Attestations from supplementary fieldwork (blue squares) on the southern fringes of the ⸢ugor⸣ ‘eel’ area help to narrow the costal discontinuity zone to barely 25 miles. Centred around Zadar, and with the identical distribution of the designations for eel and conger as on the islands facing the city (cf. maps O‑1 and O‑2), this zone may reflect the original Dalmatian Romance conditions of the former capital of Byzantine Dalmatia, with a subsequent introduction of Ven. bisoto. Alternatively, it is possible that the two ⸢ugor⸣ areas were once in direct contact, and that the coastal vernaculars around Zadar eventually adopted the type ⸢grug⸣ from nearby islands, owing to the prestige of insular fishing culture. It is worth noting that these vernaculars belong to the Southern Čakavian dialectal group that stretches all the way to Central Dalmatia and forms the very heart of the innovative ⸢ugor⸣ ‘conger’ area, while the type ⸢grug⸣ is characteristic mostly of Central Čakavian vernaculars.6 In this respect, it is interesting to observe the reactive character (cf. Telmon, 1989) of Povljana, a migrant Southern Čakavian point on the island of Pag situated in the Central Čakavian area. Exposed to the lexical pressure of the city of Pag, Povljana has endorsed ⸢grug⸣ as ‘eel’, while retaining ⸢ugor⸣ ‘conger’.
Figure 4. – Map S‑2. ⸢lokarda⸣.
10Map S‑2. ⸢lokarda⸣. Based on maps O‑3. Scomber scombrus (Q437) and O‑4. Scomber japonicus (Q438).
Figure 5. – O‑3. Scomber scombrus (Q437).
Figure 6. – O‑4. Scomber japonicus (Q438).
11From Lat. lacĕrta (ERHSJ I, s.v. argitula) or, more likely, *lacarta (Vuletić & Skračić, 2018: 109), an early DalmR. loanword, as evidenced by the treatment of Romance protonic a (felt short) > o. Attested across our network in two meanings: ‘Atlantic mackerel (Scomber scombrus)’ in the area from Eastern Istria continuing down to the borders of Northern Dalmatia, including the Kvarner islands of Cres, Krk, and (partially) Rab, and sporadically in the central part of Northern Dalmatia; ‘Atlantic chub mackerel (Scomber japonicus)’ elsewhere.
12Since the two species are very similar, Vinja (JaFa I: 280) speculates that both were designated by the same term in DalmR. vernaculars. If such polysemy ever existed in the Romance substrate, it proved intolerable for Croatian fishermen. Several lexical types appeared to designate the species not called ⸢lokarda⸣, most of them limited to a specific area, as shown by our maps O‑3 and O‑4. For Scomber japonicus: Cr. innovations ⸢plavica⸣ and ⸢ljulja⸣, and Ven. loanword ⸢lancarda⸣ < lanzardo. For Scomber scombrus: ⸢skuša⸣ of disputed etymology (cf. JE III, s.v. skuša; ERHJ II, s.v. skuša), Cr. innovation ⸢vrnut⸣, Ven. loanword ⸢škombar⸣ < sgombro, and ⸢skrdinula⸣, possibly a DalmR. loanword (cf. JE III, s.v. skrdinula).
13The area of ⸢lokarda⸣ ‘Scomber scombrus’ generally overlaps—with the exception of the island of Cres instead of Pag—with that of ⸢ugor⸣ ‘eel’, yet its borders with the ⸢lokarda⸣ ‘Scomber japonicus’ area are fuzzy. The two semantic areas are exclusive but “à limites flottantes” (Jaberg, 1935: 50), which strongly supports Vinja’s hypothesis. This becomes even more evident when we include supplementary data (squares): the microstructure reveals an important interpenetration zone of the two areas on the NW edge of Northern Dalmatia, while further corroborating the patchwork found by JAPRK in the central part of the region. The regional capital Zadar seems to have played no role in creating this intricate lexical landscape on the mainland, aligning itself, as on the map S‑1, with its islands, where *lacarta > ⸢lokarda⸣ has probably never been polysemous.
14The preceding semasiological maps served as examples of macroareal semantic differentiation: both ⸢ugor⸣ ‘eel’ and ⸢ugor⸣ ‘conger’, as just as ⸢lokarda⸣ ‘Atlantic mackerel’ and ⸢lokarda⸣ ‘Atlantic chub mackerel’, cover extensive portions of the Eastern Adriatic coast. With the following two maps, we turn to lexical types that in their semantic differentiation display a different typology, namely macroareal vs. microareal.
Figure 7. – Map S‑3. ⸢gavun⸣.
15Map S‑3. ⸢gavun⸣. Based on maps O‑5. Spicara smaris (Q425) and O‑6. Atherina hepsetus (Q449) (cf. appendix).
Figure 8. – O‑5. Spicara smaris (Q425).
Figure 9. – O‑6. Atherina hepsetus (Q449).
- 7 This is how one should interpret the lines of optimally informed A. P. Ninni (1964 [1890]: 9): “Que (...)
- 8 Note that the outermost point of the type ⸢gavun⸣ ‘Spicara smaris’ in the Kornati archipelago, whil (...)
16Contrary to Skok (ERHSJ I, s.v. agun) and Vinja (JaFa I: 392–393), clearly from Ven. agòn (ERHJ I, s.v. agun; Vuletić & Skračić, 2018: 63; cf. also Filipi, 2020), a term designating several small fish species in different Venetian dialects, foremost those of the Spicara genus, and notably male Spicara smaris.7 We distinguish this type from the co‑etymological ⸢agun⸣, registered on some Kvarner islands with historically important Venetophone centres (cf. map O‑6), phonetically more faithful to the Venetian source form. The distribution is clear‑cut: most of the JAPRK network, except for the Kvarner area, has ⸢gavun⸣ ‘Mediterranean sand smelt (Atherina hepsetus)’; a microarea corresponding to the territory of the ancient district of Šibenik has ⸢gavun⸣ ‘picarel (Spicara smaris)’.8 The supplementary microareal survey does not alter this overall picture, it only adds to the accuracy of the conclusion. The perfect correspondence between lexical and administrative borders suggests that the ⸢gavun⸣ ‘Spicara smaris’ area was formed no earlier than the 15th century, when the district of Šibenik reached its definite extent on the coast (cf. Kolanović, 1995). This is somewhat confusing, since some other designations of Spicara smaris are very old, attesting to early interest in this species: for example, ⸢gira⸣ < Lat. gĕrres is without a doubt a DalmR. loanword, as evidenced by the intact initial velar stop (ERHSJ I, s.v. gara; JaFa I: 346).
17According to the principle of geographic coherence (cf. Telmon, 1989), the Northern Dalmatian insular microarea of ⸢gira⸣ can be interpreted as a remnant of a much larger area of this lexical type, nowadays characteristic of Central and Southern Dalmatia (cf. map O‑5). Therefore, some external factor—most likely the introduction of Venetian designations for the fish of the Spicara genus—must have reshaped the original lexical landscape of Northern Dalmatia, with the Zadar area accepting ménola > ⸢manula⸣, and Šibenik area adopting agòn > ⸢gavun⸣, both for Spicara smaris. In this respect, Šibenik shows remarkable originality and vitality (in the sense of Terracini’s writings; cf. also Telmon, 1989) against the rest of Dalmatia that selected agòn > ⸢gavun⸣ for Atherina hepsetus. Moreover, it goes a step further by adapting Gr. βρέφος (JaFa I: 397–398; JE I, s.v. brfun) to the morphological model offered by ⸢gavun⸣ (Vuletić & Skračić, 2018: 44), thereby creating ⸢brfun⸣ ‘Atherina hepsetus’ (cf. map O‑6).
Figure 10. – Map S‑4. ⸢pauk⸣.
- 9 Vinja’s experience and our preparatory fieldwork have shown that the lexical distinction between di (...)
18Map S‑4. ⸢pauk⸣. Based on maps O‑7. Trachinidae (Q435) and O‑8. Uranoscopus scaber (Q436).9
Figure 11. – O‑7. Trachinidae (Q435).
Figure 12. – O‑8. Uranoscopus scaber (Q436).
19An obvious metaphor, connecting a fish capable of inflicting painful sting with a spider (pauk, cf. JaFa I: 137). The compact area of ⸢pauk⸣ ‘weever’ covers most of Central Dalmatia and the SE portion of Northern Dalmatia, reaching the outskirts of Zadar, where it meets the area of ⸢ranj⸣ < Ven. ragno (cf. map O‑7). As is the case with ⸢ugor⸣ ‘conger’ (cf. map S‑1), its NW limit falls very close to the Southern Čakavian – Central Čakavian dialect border.
20In the SE, a contiguous microarea has ⸢pauk⸣ ‘stargazer’. This fish is equipped with two venomous spikes on its gill covers, so the motivation is essentially the same as with ⸢pauk⸣ ‘weever’. We might be dealing not with a case of semantic differentiation but with the results of two independent denomination processes relying on the same metaphor, however, the geographic contiguity of the two areas warns against that possibility. In fact, the area of ⸢pauk⸣ ‘stargazer’ coincides with the one where the designations of weever belong to the type ⸢dragana⸣ < Gr. δράκαινα, an old epichoric loanword probably mediated by some early DalmR. variety, further modified by the euphemistic intervention of Cr. drag ‘dear’ (cf. Bazina 2024: 140). It is easily conceivable that the westernmost vernaculars of this area, while not participating in the innovation of the neighbouring area, borrowed one of its fish names, fitting the needs of the local lexical system from a motivational point of view.
21As far as methodology is concerned, this map is a good example of the importance of complementary macroareal and microareal geolinguistic research. The JAPRK network, although seemingly quite dense in this area, might lead one to believe that the innovation ⸢pauk⸣ ‘stargazer’ is restricted to a tiny zone where the island of Hvar approaches the mainland and where both locations attesting this type, Brist (on the mainland) and Sućuraj (on Hvar), belong to the Štokavian macrodialect. Furthermore, since the population of Sućuraj came from the mainland, it is reasonable to assume that the innovation originated there. However, the microstructure shows not only that two apparently isolated Štokavian points participate in a broader microarea, but also that the centre of that area is formed by Čakavian vernaculars of the island of Korčula. Furthermore, it is worth noting that the five rural points on Korčula attesting ⸢pauk⸣ ‘stargazer’ show remarkable solidarity against two small urban centres on two opposite sides of the island, Korčula and Vela Luka, which have the type ⸢bežmek⸣ ‘stargazer’. Since the latter type is well documented in the Dubrovnik area, further south (cf. map O‑8), there are solid grounds to believe that it reflects a more recent contact with the Southern Dalmatian capital (as is the case with ⸢ovrata⸣ ‘gilt‑head bream, Sparus aurata’, cf. Bazina, 2024: 111).
Figure 13. – Map S‑5. ⸢pizdica⸣.
22Map S‑5. ⸢pizdica⸣. Based on the data for as many as eleven concepts found in the JAPRK questionnaire, and several referents, mostly molluscs, not present in the questionnaire but mentioned by informants during our fieldwork. Situations where the informants warned us that a certain sea animal we did not ask about was named pizdica were not rare. Learning from this experience, we eventually introduced a supplementary semasiological question regarding pizdica.
- 10 And vice versa: shell names not based on this association often serve as euphemistic designations f (...)
23Unfortunately, this intervention did not affect all our surveys, so the present map is necessarily incomplete. Given these circumstances, we provide no onomasiological maps for ⸢pizdica⸣ here. A transparent type, it represents a diminutive form of pizda, colloquial and synchronically vulgar term for female genitalia. The association of open shells, particularly of those Bivalvia species with longitudinal bilaterally symmetrical meat, with this part of human anatomy is probably as old as our naming of the world.10 The present map is a good illustration of how persistent such motivation can be in folk taxonomy. Nonetheless, the apparent semantic mess on the map shows that it is also subject to changing social scruples and restrictions. That is, the actual geographical dispersion of ⸢pizdica⸣ together with the complexity of its referential domain likely reflects what Jaberg (1936: 43, 47) called “type lexicologique décadent” or “type lexicologique en recul”. The historical reasons for a “retreat” of a wordform considered obscene are easy to list, from common social restraints in a Christian cultural setting to the ever‑growing access to formal education, and, in the case of seafood, a necessity to reach broader markets. What concerns us here is the geographical and denotational pattern of that retreat, which in the case of ⸢pizdica⸣, turns out to be quite interesting.
- 11 Rare examples of polysemy are found on the island of Čiovo, where ⸢pizdica⸣ denotes both Mytilus ga (...)
- 12 Cf. Vinja’s interpretation of the motivation of pidzica ‘Naria spurca’ (JaFa II: 142).
24Let us first observe the absence of this lexical type in Southern Dalmatia: either it was never present in this area or social restraints in the ancient Republic of Dubrovnik were more efficient that elsewhere. On the opposite side of our network, the entire Kvarner Bay must have known this lexical type, as isolated attestations in that area seem to suggest. Second, the most frequent referent of ⸢pizdica⸣ is, according to our data, the Mediterranean mussel, Mytilus galloprovincialis. This is hardly surprising, given the (internal) anatomy of this highly valued mollusc. The JAPRK network shows an important area of this subtype in Central Dalmatia, involving urban centres of Split, Trogir, and Hvar, with scattered attestations further to the NW (black circles). The microstructure (black squares), on the other hand, reveals that the Central Dalmatian area of ⸢pizdica⸣ ‘Mytilus galloprovincialis’ is quite dense, particularly along the coast, but it also identifies important remnants of this subtype in the Zadar archipelago, a geolinguistic structure that the macrostructure fails to document.11 Modiolus barbatus, the bearded horse mussel, is quite similar, therefore, the attestations of ⸢pizdica⸣ denoting this mollusc in Northern Dalmatia (orange) can be attached to the area of ⸢pizdica⸣ ‘Mytilus galloprovincialis’. Whether it is hazardous to suppose that this is the primary meaning of ⸢pizdica⸣ depends on the interpretation of many other meanings documented in our research. We cannot reasonably exclude the scenario of initial polysemy.12
25One could even imagine that ⸢pizdica⸣ was once the designation par excellence for seashells, except perhaps in Southern Dalmatia, probably denoting the most common or most valued species in each specific ecoregion. Nevertheless, going back to Jaberg’s teaching on dispersed semantic areas of retreating lexical types, we hypothesize that initially ⸢pizdica⸣ could have had one meaning, probably ‘Mytilus galloprovincialis’, in most of the Croatian Adriatic. Facing growing social restraints, this lexical type would have “pulled back”, so to say, to less valued referents, rarely named outside the local context, some of those not even remotely recalling the initial motivational impulse. This process of extreme semantic differentiation would have followed the pattern of local innovations, occasionally giving rise to relatively compact semantic microareas, such as ⸢pizdica⸣ ‘Cerastoderma edule’ on the island of Pašman (light blue), or ⸢pizdica⸣ ‘Naria spurca’ on the northwesternmost Zadar islands (magenta).
26In the chapter “Methodological remarks”, we presented two primary objectives of this paper: from the methodological perspective, a brief discussion of the process of establishing a survey network, i.e. the implications encountered after the selection of survey points; and from the analytical point of view, the geolinguistic interpretation of several semasiological maps, following Jaberg’s heuristic principles on the semantic differentiation of lexical phenomena. Establishing a network represents a particularly demanding phase of the research, where shortcomings are apparently inevitable. An excessively dense survey network would certainly raise many challenging questions, particularly when it comes to lexical types confined to very narrow areas (i.e. geolinguistic noise).
27On the other hand, a thin raster can lead to a more balanced map, but it undoubtedly also hinders crucial attestations, especially when dealing with the historical dimension of geolinguistic structures. This problem was amended by introducing the combined basemap, which integrates data from multiple networks of different density, yet mostly collected using the same set of questions on the lexicon of marine fauna. Microstructural data were collected through the “point par point” fieldwork method, an idea taken from traditional linguistic geography. The combined basemap offers both a broad insight into the spatial distribution of different semantic areas associated with the same word form and the potential to retrieve data as detailed as needed. The result, we believe, is a precise analysis of geolinguistic macrostructures and microstructures, which are clearly observable at different levels of network density. Semasiological maps have already proved useful in revealing linguistic and cultural behavioural patterns of basilects (cf. Jaberg, 1936). However, even though Jaberg’s innovative ideas were mostly met with approval (cf. e.g. Bruneau, 1937; Malkiel, 2016), it seems that modern linguistic geography is more concerned with the application of exact quantitative methods, which often lack proper philological analysis, than with the interpretation of individual onomasiological and semasiological maps, especially when it comes to the analysis of geolinguistic structures. This contribution attempts to partially fill this gap.
28We have discussed several semasiological and onomasiological maps, trying to shed light on the dynamics of geolinguistic structures in the Eastern Adriatic, as reflected in the spatial distribution of the lexicon of marine fauna. The first two maps show exclusive semantic macroareas of different meanings of the ⸢ugor⸣ (cf. map S‑1) and ⸢lokarda⸣ (cf. map S‑2) types, with an almost identical spatial pattern. In the case of the ⸢ugor⸣ lexical type, Central Čakavian points in Northern Dalmatia represent a strong buffer zone between two mutually exclusive macroareas. However, as the accompanying onomasiological maps show (cf. maps O‑1 to O‑4), Northern Dalmatia is an area of important lexical dynamics, especially when it comes to the relationship between mainland and island points, manifested, on the one hand, in the conservation of the ⸢grug⸣ or ⸢skrdinula⸣ types, and on the other, in the acquisition of Venetian loanwords (e.g. bisoto and lanzarda). It is perhaps not surprising that in this colourful geolinguistic context, smaller areas emerge with their reactive character, often a non‑negligible feature of migrant points (cf. Bazina, 2024). This is the case with Povljana, where a specific semantic shift between the ⸢ugor⸣ and ⸢grug⸣ types has occurred (cf. map S‑1). Two maps (S‑3 and S‑4) deal with the semantic differentiation between macroareas and microareas. In the case of the map S‑3 and following onomasiological maps, it is shown that the original lexical landscape in Northern Dalmatia has changed mainly because of the introduction of various Venetian names for fish of the Spicara genus (e.g. ménola and agòn). In addition, the discussion of map O‑6, where points belonging to the old Šibenik district created the ⸢brfun⸣ ‘Atherina hepsetus’ type, leads us to the conclusion that morphological phenomena frequently indicate remarkable linguistic vitality (cf. Terracini, 1964). While these conclusions are based on the JAPRK’s macro‑network, supplementary material certainly contributes to a better understanding of the history of ⸢pauk⸣ ‘stargazer’ in Central Dalmatia (cf. map S‑4). The last semasiological map (S‑5) refers to Jaberg’s idea of “type lexicologique décadent” or “type lexicologique en recul”, illustrated both by the large number of different semantic values and their intricate geographical distribution. This so‑called accident sémantique is easily attributed to the affective stimulus of the word form pizdica. Our interpretation, based on microstructural attestations, marks just one possible explanation of the extreme semantic differentiation of ⸢pizdica⸣, but the history of this lexical type remains rather elusive.