Swedifying the Other

The Large-Scale Cadastral Maps and Place-Name Policy Making

The most ambitious large-scale cadastral mapping project of the seventeenth century was initiated on April 4, 1628, in Sweden. Gustav II Adolf instructed the state’s Generalmatematikus Anders Bure to establish a central land survey function to survey and map the conglomerate state, thus setting an unprecedented economic and political goal (Samlingar 1, no. 1 [Samlingar i landtmäteri. Saml. 1, Instruktioner och bref, 1628–1699, 1901, Stockholm]; Helmfrid 1959; Kain and Baigent 1992, s. 49–75; Baigent 2005; Höglund 2013a, 2013b; Tollin 2021, s. 61–113, 369–86; Petrulevich 2022).¹ Obviously, foreign role models played an important role with respect to choices of surveying techniques and map design. However, the centralized land survey function, the systematic survey work, and the sheer scope of the endeavor single out this Swedish project that lacks a comparable counterpart in Europe or the world. The project’s outcome, more than 26,000 large-scale cadastral maps, constitute a unique source material. These can be divided into about 12,000 so-called older cadastral maps, produced between 1620s and 1650s, and about 14,000 so-called younger cadastral maps dated to the last decades of the seventeenth century, from 1680s until 1700. These two categories differ in many respects, although the technological underpinnings, including surveying techniques, remained mostly the same. Most important, the relative homogeneity of map design gradually increased toward the end of the period due to centralized governing and detailed directives minutely regulating the surveying and mapping process (Petrulevich forthcoming a).²

The three map volumes chosen for this study, Å1, A 1b, and BI, represent final, redrawn products of the tedious and time-consuming surveying and mapping work. Unfortunately, the land surveyor instructions never describe surveying, measuring, or mapping techniques themselves, although some related specific requirements (e.g., to use the same scale or the same type of instruments) do occur (Petrulevich forthcoming a). The surveying and mapping process can be reconstructed in the following way (the outline is based on Rålamb 1690; Samlingar 1 no. 5, 10; Svärdson 1928, 144–47; Tollin 2021, 76–91). Necessary fieldwork to survey the area and draft a map was likely conducted mostly in spring and summer, but even in autumn and winter if the weather allowed. The survey itself can be described as a form of scale drawing centered on the use of a plane table in the field. Representations of mapped objects such as fields or meadows were simplified to polygons drawn in true-to-scale-fashion with the help of a diopter, an early form of alidade. These first drafts, called “concept” maps (koncept), were later redrawn and colored to produce final outputs, so-called renovated maps (renovationer). To the best of my knowledge, no draft maps of the Å1, A 1b, and BI volumes have been preserved. According to the map signatures, Å1 was drawn by one responsible surveyor, Erik Eriksson Vidman, A 1b by the ordinary surveyor Olof Mört (d. 1703) and the extraordinary surveyor Magnus Bergman, and BI by four different land surveyors, individually responsible for different parts of the Wolgast district, Simon Skragge (d. 1703), Arvid Carlmark (1665–1712), Brynolf Hesselgren (1666–1734), and Olof (Olaus) Spaak (1668–1719).³ It is likely that many more people partook in surveying and mapping, because the plane table surveying required at least two people. Earlier scholarship assumes that land surveyors, their disciples or apprentices, locals, and even family members could have been involved.

Apart from the mapped features and symbolic representations of other objects, the maps often include place-names and more or less detailed descriptions of the settlements and homesteads, their inhabitants and assets, so-called notarum explicatio. In the Pomeranian material, the descriptions corresponding to notarum explicatio were gathered in separate volumes, Urschriften (e.g., Drolshagen 1920, 91–100; Höglund 2013a, 113–14; Petrulevich forthcoming a).⁴ These notes are in many cases exceptionally detailed and provide much more information than their counterparts in the Finnish and especially the older, Swedish material. Many of the place-names, especially settlement names, included in the maps and notarum explicatio or separate descriptions were most likely retrieved from older cadastral sources (cf. Petrulevich 2022, 108f.). However, microtoponyms (i.e. the names of minor features in the landscape such as fields, meadows, pastures, forests, hills, hollows, marshes, border markers, minor lakes or brooks⁵) in most cases appear in the maps for the very first time.⁶ These microtoponyms were collected by land surveyors from multiple informants in situ. The rich Pomeranian description material offers many accounts confirming this, for example, “Effter deße åboernes ord och utsagu är detta dorff af fölliande beskaffenheet” [According to the words and statements of these inhabitants, this village has following characteristics], “Ett sankt alekiärr kallas Demse Mor” [A marsh called Demse Moor], and “dhe andra sanka alekiärr hafwa inga namn” [the other marshes have no names] (Urschrift Bd 35, U38, U143; cf. Petrulevich, forthcoming a). Thus, the uniqueness spectrum of the Swedish cadastral maps and accompanying descriptions includes a clear onomastic dimension.

In two previous studies (Petrulevich 2022, forthcoming a) I have investigated the body of directives and instructions explicitly targeting naming of mapped features. Two of these are of primary importance in this article: Kammarkollegiet’s instruction on surveying Pomerania from August 15, 1691 (no. 103, Samlingar 1, SE/RA/420571/03/F4/1), and director of the Land Survey’s clarification on surveying towns from March 30, 1697 (no. 175, Samlingar 1, SE/RA/420571/03/B/B1/8). The directives explicitly address processing place-names in multilingual areas and use the term Swedify, cf. förswänskat [Swedified] and med rätt förswänskade ord [in correct Swedified words] (see next section for the concept of Swedification). The first one requires Swedification of names (nomina propria) and appellatives (appellativa)⁷ in High and Low German:

Wijd alle sine Mätningar och úti alle sine annotationer och beskrifningar, skolla dhe till bättre únderrättelse, så för Oß som posteriteten behålla behålla och úpföra alla nomina propria och appellativa úppå alt, hwad únder dheras Embetz förrättningar förekomma kan på det tÿska eller rätta giemena Pommerska Språket, och dher hoos hwart och ett effter des natúr och egenskap, igenom nu särledes accúrat in=hemptad únderrättelse och rätt förstånd, för=swänskat tillsättia och således sakerne rätt explicera.

(When surveying and measuring as well as in their annotations and descriptions, they [land surveyors] should keep and put down—for better understanding of ours and of the coming generations—nomina propria and appellativa of everything that can appear when they carry out their duties in the German or the common Pomeranian language [Low German]; but also add each and every one of them in a Swedified form in accordance with their nature and characteristics by making accurate inquiries, and thus explicate things in the correct fashion.)

The second is concerned with the treatment of names, appellatives, and descriptions of relevant mapped features in many more languages, Finnish, Estonian, Latvian, common Livonian, High and Low German, as well as Northern Low Saxon:

Slútligen måste . . . och så úti dee finska, est= lät=tiske, gemene Lijfländske och tÿska, ge=mene Pommerske, samt Brehmiske-Språken, antecknas effter nödwändig=heeten, icke allenast nomina propria och appellativa på lägenheeterne, samt deras jordemoners betÿdelße med rätt förswänskade ord jämwäl tillsatte.

(Finally . . . when necessary, not only nomina propria and appellativa of the localities have to be noted in the Finnish, Estonian, Latvian, common Livonian and German, common Pomeranian [Low German] and Bremian [Northern Low Saxon] languages, but also the significance of their type of soil and farming etc. likewise must be added in correct Swedified words.)

In this case, the link between Swedification and name processing appears to be indirect and subject to the interpretation of the clause as a whole, rather than its individual constituents. My interpretation of the 1697 directive is somewhat affected by the influential Samlingar edition (1901) of the land surveyor instructions abundantly cited in earlier literature. There, Swedification is more explicitly linked to all the three relevant entities, names, common words and the name referents’ descriptions. See the edition’s interpretation “På de Finska, Estniska, Lettiska, gemena Livländska och Tyska, gemena Pommerska samt Bremiska språken antecknas efter nödvändigheten nomina propria och appellativa på lägenheterna samt deras jordmåners art och bruk etc., hvarvid dock jämväl tillsättes deras betydelse med rätta försvenskade ord” (no 175, Samlingar 1, 232) [When necessary, nomina propria and appellativa of the localities as well as their type of soil and farming etc. should be noted in the Finnish, Estonian, Latvian, common Livonian and German, common Pomeranian (Low German) and Bremian (Northern Low Saxon) languages, but their meaning likewise should be added in correct Swedified words].⁸

It is also important to note that the regulations and directives from the middle and the end of the seventeenth century likely confirmed and reinforced the more or less naturally evolved, already established practices to ensure the uniformity of the surveying process and its outputs (Petrulevich forthcoming a). For example, the explicit demand to use older cadastral records in the 1643 instruction most likely indicates that these records were used from the beginning (Tollin 2021, 96)—possibly (with some exceptions) the instruction aimed to reinforce this. Similarly, the 1691 and 1697 directives dealing with non-Scandinavian names in the Swedish conglomerate state can be seen as traces of already established practices the senders of the documents sought to reinforce.⁹ Thus I do not assume that the directives in question introduced new strategies of collecting or processing name materials and, consequently, that the practices changed after the regulations entered into force. Finally, the existence of or amount of directives does not necessarily imply or guarantee immediate obedience on behalf of land surveyors and thus uniform outputs, maps, and descriptions.

Place-Name Swedification and Contact Onomastics

What does Swedification with respect to proper names, descriptions of names’ referents and common words actually mean? The past participle försvenskat (förswänskat and förswänskade ord in the directives) corresponds to the Swedish verb försvenska attested as early as 1575 with the meaning “översätta l. överflytta (ngt) till svenska; kläda (ngt) i svensk (språk)dräkt” (SAOB/försvenska, = för-svenska, II 3 c 1, https://svenska.se/saob/?id=F _3169-0113.7Bw3&pz=7) [translate or transfer (smth) into Swedish]. In other words, the land surveyors are asked to translate non-Scandinavian names, descriptions of names’ referents, and common words into Swedish. While translating a phrase or a word from one language to another appears to be a straightforward task, translating names is not. Translation belongs to the core repertoire of place-name adaptation strategies in language contact situations (Sandnes 2016 with references, see also below), but names do not necessarily have any transparent lexical meaning to be expressed in Swedish or any other recipient language (cf. Christensen and Kousgård Sørensen 1972, 9–26; Coates 2016; Nyström 2016).

For this reason, my analytical framework is expanded to combine the insights from interpreting the key directives and from earlier research on adaptation strategies of borrowed or replicated place-names in contact onomastics. The rationale behind this expansion lies in the ambition to collect all possible traces of adaptation to Swedish and thus gain a comprehensive understanding of how non-Scandinavian place-names in three language contact situations were processed. For analytical purposes, the concept of Swedification is thus understood broadly as linguistic adaptation to Swedish. However, a conscious bias to focus on possible strategies aimed to express lexical meaning is introduced to match the requirements and phrasings of the seventeenth-century directives.

Contact onomastics investigates how names are processed in language contact situations (Sandnes 2016). One of its major goals is to establish the factors that constrain contact-induced change in name data sets, an ambition inspired by the mother field of contact linguistics (Thomason 2012; Grant 2020a). Linguistic analysis of so-called interference or transfer phenomena (i.e. the influence of the recipient language on linguistic aspects of items borrowed from a donor language) is a point of departure for a discussion on explanatory factors inspired by contact linguistics (Thomason and Kaufmann 1988; Thomason 2001; Winford 2020). The bulk of name adaptation models are based on structuralist analysis of interference phenomena or adaptation strategies on different language levels, for example, phonological, morphological, or lexical (Sandnes 2016; Pedersen 2015; cf. Weinreich 1953; Grant 2020b, part I). To the best of my knowledge, linguistic inquiry into the name data from the Swedish large-scale cadastral maps has been rather modest. In most cases, the maps and descriptions provide raw name data for noncontact linguistic analyses (Digitales Flurnamenbuch Vorpommern, https:// flurnamenbuch-vorpommern.germanistik.uni-greifswald.de; Neumüller 2013; Pihl 2014; Vollmer 2021; Kepsu 2023). There are thus no comparable studies of place-name processing in the context of the Swedish seventeenthcentury surveying and mapping in different language contact areas.

Linguistic adaptation of place-names in the broadest sense includes “‘all the changes that a replicated name can undergo in the recipient language both on replication and further on at all language levels including both addition and omission of linguistic elements’” (Petrulevich 2016, 49, my translation). As a rule, the changes in question comply with the linguistic system of the recipient language. I distinguish between the following adaptation strategies, also evident in earlier scholarship under the same or similar labels (Sandnes 2003, chap. 9, 2016; Pedersen 2015; Petrulevich 2016, 53–60): phonological adaptation, morphological adaptation, lexical adaptation, onomastic adaptation, semantic adaptation or translation, and epexegetic adaptation. Of these, translation is the only adaptation type that requires some basic knowledge of the donor or the recipient language—depending on who is responsible for the adaptation in question, for example, the land surveyor or the land surveyor’s informant.

Phonological adaptation implies that the sounds and sound combinations of the replicated place-name that are not compatible with the phonotactics of the recipient language are substituted by other, compatible sounds and sound combinations (cf. Petrulevich 2016, 57f.). In the language contact situations involving two or more closely related languages, such as Finnish and any of the Sámi languages or Swedish and Low German, substitutions originating at the phonological level can lead to de facto translations. This phenomenon is called etymological nativization in some earlier literature (Aikio 2007, 164f.; Heikkilä 2014, 150f.). In this article, all examples of semantic adaptation or translation are analysed as such, even if the majority of the translation examples in the Pomeranian material can be seen as the results of etymological nativisation, for example, Swedish hed [heath], dialectally also “woodland,” corresponding to Low German Heide [woodland]. It is also necessary to point out that a very strict definition of phonological adaptation is employed here because of extensive orthographic variation. Thus, only simplifications of diphthongs and consonant clusters are considered. For instance, the Finnish island name Töyräluoto, built to töyrä or töyry (SES/töyrä, https://kaino.kotus.fi/suomenetymologinensanakirja/) [hummock, hillock, bank] and luoto [islet, rock], is spelled as Töareloto (A 1b:3), that is, the [uo̯  ] diphthong is simplified to a monophthong. The Finnish settlement name Hirvijärvi, originally a compound lake name, and the lake name Hirvisalmi, built to hirvi [elk, moose], järvi [lake] and salmi [strait, sound], appear as Herijerwi and Hijrisalmj respectively (A 1b, 19), that is, the [rʋ] consonant cluster is simplified. However, such cases as Kiffwijsalme for Kivisalmi (A 1b, 10), containing Fin. kivi [stone] and salmi [strait, sound], Kamppikaskj for Kamppikoski (A 1b, 23) and Koitagenkoskj for presumably Koittojenkoski (A 1b, 16), containing Fin. koski [rapids] and possibly Fin. kamppi [tripping] and koitto [dawning] in oblique form, respectively, are not considered. Here, the spellings <e> for <i> in -salme, <a> for <o> in -kaskj and Koitagen-, and <g> for <j> in Koitagen- are considered orthographic variants.

By morphological adaptation I mean such cases when morphological elements in the donor language are reinterpreted as morphological elements in the recipient language (cf. Petrulevich 2016, 58) or when morphological elements of the recipient language are added to the replicated place-name. For example, the Finnish border marker name Kuninkaankallio, built to kuningas [king] in genitive and kallio [rock], is rendered as Kuningenkallio (A 1b, 35), that is, the Finnish morpheme is substituted by the Swedish -ing and -en. The German Sommerfeld appears as Sommerfeltet in Swedish (BI, 6); the Swedish definite article -et is added to the name.

Lexical adaptation means that the replicated place-name as a whole or any of its constituents are substituted with words of the recipient language that have a phonological form similar to that of the replicated name (cf. Petrulevich 2016, 58). For instance, the generic of the originally Finnish compound water name Hortlax, built to hurtta [(large) dog, hound] and laksi [bay, inlet] (cf. Pellijeff 1988, 34f.), attested as a settlement name in the seventeenth-century material (Å1, 71), was possibly interpreted as Swedish lax [salmon]. The Swedish rendering of the Finnish field name Alimmainenvainio, built to alimmainen [undermost] and vainio [field], Alnamainenwainio, possibly exhibits reinterpretation of the onset of the name as the Swedish aln [historic measure of length] or alna [measure in alns] (A 1b, 1).

Onomastic adaptation covers those cases when the replicated place-name as a whole or any of its parts are reinterpreted as names or name constituents of the recipient language that phonologically resemble the replicated name (cf. Petrulevich 2016, 58). For instance, the West Slavic settlement name Wangelkow transmitted to Swedish via German appears as Wrangelkow (BI, 12), that is, the onset of the name is reinterpreted as the family name Wrangel. The Wrangel family had a close relation to Pomerania and the Wolgast province, where they owned a castle known under the name Wrangelsburg from 1650s (earlier Vorwerk).

Semantic adaptation or translation implies that the lexical meaning of the replicated name is expressed in the recipient language. For instance, the Finnish field name Lounasvainio built to lounas [Southwest/ern] and vainio [field] was translated to Swedish as Sydvästra åkern (A 1b, 2), while the German Wrangelsburgs Heide is in turn rendered as Wrangelsburgsheden in Swedish (BI, 40, Urschrift Bd 35, U532).

A replicated name is seen as adapted epexegetically if an epexegetic name element, that is, a name element specifying the name’s type of referent, from the recipient language is added to the name (cf. Petrulevich 2016, 59). The Finnish settlement name Laiska, attested as Laiskahemman in A 1b, 2, and the Finnish marsh name Kovastenkangas, rendered as Kovastenkangashed in A 1b, 21, are examples of epexegetic adaptation, because the Swedish epexegetic name elements hemman [homestead] and hed [moor, marsh] respectively have been added to the names. Epexegesis is seen as an adaptation strategy here, but the phenomenon can also be seen as a strategy to form new native names (cf. discussion in Petrulevich 2016, 51f.).

Apart from these established place-name adaptation strategies, I introduce two more, explication and independent rendering, to properly observe all possible Swedification strategies in the seventeenth-century context. By explication I mean a description or specification of the name referent, usually in the same noun phrase. For example, in the prepositional phrase “wed engen Gregerort” [near the meadow Gregerort] in the Pomeranian material (BI, 7; Urschrift Bd 35, U55), the referent of the name Gregerort is specified as Swedish äng [meadow]. This strategy can be seen as a counterpart to epexegesis. The difference between the strategies lies in the syntactic position and status of the specifying element (the head in the noun phrase in explication, ängen Gregerort, and an epexegetic generic in the meadow name *Gregerortängen in epexegesis). Independent rendering covers those cases when land surveyors are assumed to have coined their own names for landscape features and other referents to replace the already existing non-Scandinavian names (cf. Pitkänen 1997). For instance, the field names built to words denoting cardinal direction such as Det norra fältet [the northern field] in the Pomeranian material (BI, 25; Urschrift Bd 35, U298) are seen as coined by land surveyors.¹⁰ Also, the Swedish name of the boundary mark Horvarået built to horva [small enclosed field or meadow] and rå [boundary mark] is seen as an independent rendering by a responsible land surveyor because the feature in question bears a Finnish name, Humalatarha, cf. humala [hop] and tarha [corral, yard], according to the same map (A 1b, 1).

The foregoing adaptation strategies can be combined in the attestation material. In such cases, the place-name attestation is analyzed as an example of two or more adaptation strategies. For instance, the Low German field name Brakfeld built to Brak [fallow] and Feld [field] appears as the adapted Brakfeltet in the phrase “Brakfeltet el:r trädet.” The attestation is labeled as an example of both morphological adaptation and explication because the name is provided with the Swedish definite article and the name referent is specified.

There are several main internal and external factors that can affect the linguistic outcomes of language contact or contact-induced change (Thomason & Kaufman 1988, chap. 3; Thomason 2012, 39; Winford 2020): typological similarities between the contact languages, the social status of the groups in contact, the intensity of the contact, the degree of multilingualism in the population, and if the language-contact situation leads to language maintenance or language shift. In contact onomastics, the discussion of the constraints on place-name replication and adaptation follows these generalized observations (Sandnes 2003, 2016; Pedersen 2015; Petrulevich 2016, 35f.). In those transmission situations where place-names are assumed to have been handed down by particular people such as scribes or land surveyors via written sources or maps, the language user’s perspective, their needs, competence, and attitudes, gain in significance; “speakers’ attitudes can trump expectations for types and degree of both externally and internally motivated change” (Thomason 2012, 46) and “the choice of adaptation strategy therefore lies in the domain of the name user, while its resulting form is in most cases governed by the linguistic system of the target language” (Petrulevich 2016, 263).

The three language contact areas under study, Västerbotten, Turku and Pori, and Pomerania, have different characteristics when it comes to language typology, social status, intensity of the contact, and degree of multilingualism. The two major common denominators are the language maintenance contact scenario, that is, such contact situations where language users keep their respective language or both languages, and the prominent role of the Swedish land surveyors. The relatively short temporal window for the surveying and mapping missions and thus any possible language contact between the surveyors and the local population is also of much importance. Based on the map evidence, it is assumed the surveying and mapping work took six to seven years (Västerbotten and Pomerania, respectively) to just over a decade (Turku and Pori). In Västerbotten, the Sámi and Finnish place-name material consists exclusively of settlement names, for example, Hortlax, and river names, for example, Lule älv. These often represent examples of name loan or secondary name formations built to names or landscape features replicated and adapted long before the surveying and mapping took place (see discussion below). For this reason, the names in question constitute substrate names in the seventeenth-century Swedish context. Any contacts Erik Eriksson Vidman might have had with the Sámi- or Finnish-speaking population have not left any imprint on the maps.

Earlier scholarship into the nature and circumstances of the Finnish– Swedish language contact in Finland, where the speakers of both languages resided since at least the Middle Ages, is extensive (to name just a few: Wallén 1932; Thylin-Klaus 2019; Kuvaja 2019 with references). In the present article, the issue of possibly shifting language borders between Swedish and Finnish in the historically Swedish-speaking Turku and Pori is of minor importance because the collected place-name evidence itself, especially the many Finnish microtoponyms, are a clear indication that Finnish was the language of the peasants in the studied area in the seventeenth century.¹¹ The most important issue is instead to establish if the Swedish land surveyors, Olof Mört and Magnus Bergman—or their possible informants—were bilingual to some degree. Unfortunately, in principle it is impossible to draw any safe conclusions about the language competence issue in this case. Both land surveyors are poorly attested, but it is likely that at least Bergman was Sweden-born (Petrulevich forthcoming a). It is also reasonable that the land surveyors developed some proficiency in Finnish over the years and that the locals they came in contact with could have been bilingual.

Finally, in Pomerania there is no extensive prehistory of intensive language contact between Swedish and High and Low German, as opposed to the major influence the German varieties exercised on the Swedish language. The period of Swedish governing in Pomerania (1648–1815) was relatively short and did not leave any imprint on the composition of the peasants in the area. The four responsible land surveyors, Simon Skragge, Arvid Carlmark, Brynolf Hesselgren, and Olof Spaak, were born and raised in Sweden. Judging by extensive code-switching to German in Pomeranian land surveyor descriptions, all of them had good or excellent command of German, not least due to university studies in Uppsala (Petrulevich forthcoming a and b). In any case, German was the most widely spread second language in Sweden at the time (cf. Andersson 2011).

Swedifying the Other: General Tendencies and Individual Traits

From Place-Name Collection to Place-Name Analysis: Some Notes on Material and Method

The place-name attestation data extracted from the three map volumes (Å1, MHA A 1b, and BI), presented initially has been grouped into five data sets: Västerbotten data, Turku and Pori data, Pomeranian data excerpted from BI maps, Pomeranian data excerpted from the description volume Urschrift Bd 35, and the combined Pomeranian data. This approach is motivated by the issue of comparability posed by the historical maps and place-name materials. As mentioned earlier, the younger large-scale cadastral maps (here, the Finnish and Pomeranian material) exhibit more homogeneity with respect to design and implementation compared with Västerbotten maps, although the surveying and mapping techniques were stable throughout the seventeenth century. For instance, the inclination to collect and represent microtoponyms was higher in land surveyors in Turku and Pori and Pomerania than in Västerbotten (Figure 1).¹²

The three language contact situations underpinning the data sets differ significantly. Once again, the older, Västerbotten data is the odd one out. The place-name material of Sámi and Finnish origin is limited to settlement names (originally nature names in most cases) borrowed and in some cases attested centuries earlier. Thus, the language contact represented by the place-names in question can be seen as interrupted and the attested Sámi and Finnish lexemes and place-name elements as substrate units embedded in the overall Swedish context. In contrast, we are dealing with ongoing language contact between Finnish and Swedish in Turku and Pori and German and Swedish in Pomerania. These contact situations are not entirely equivalent either. Finland was a part of the Swedish state and a home to its Swedish-speaking population in coastal areas and beyond since the Middle Ages. Pomerania had the status of a dominion and was obtained by Sweden as a result of warfare as late as the first half of the seventeenth century. In addition, the Pomeranian toponymy showcases a considerable amount of West Slavic place-names, mainly settlement names, used in the monolingual German-speaking context. Despite these differences, the Finnish and the Pomeranian data sets can be seen as comparable with respect to their contact linguistic background of continuous language contact.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Amount of attestations in total and by type of referent across the data sets.

The totality of the Pomeranian data is divided into two subsets—map data and description data—to better match the Turku and Pori data set. The rationale behind this approach lies in the ambition to offer more nuanced comparisons between the Finnish and the Pomeranian data. The Pomeranian description volume Urschrift Bd 35 contains detailed information and thousands of name attestations with no equivalent in the Finnish map volume. At the same time, the Pomeranian maps do not provide any description information comparable to the Finnish notarum explicatio.

What counts as a place-name or a place-name attestation in a map? There is no straightforward answer in all cases. Settlement names and names of administrative units characterized by continuity of transmission and usage are the easiest to delimit but are of least interest for this investigation because these often appear in older cadastral records presumably used by land surveyors. However, the Turku and Pori maps contain names of minor homesteads with seemingly no previous cadastral record history, for example, Heikkilähemmanet [Heikkilä/Heikkilä] and Nikulahemmanet.¹³ When it comes to microtoponyms, it is much more difficult to maintain any clearcut borders to the rest of the lexicon because of lack of context. Ordinary lexemes such as lake or meadow or field can function as names in speech. For this reason, I have chosen a very broad definition of what a name in a map can be—any linguistic item or expression denoting a mapped landscape feature— and have accordingly collected items that belong to onomastic borderlands. According to one guiding principle, any lexemes dealing with the quality of land such as the Swedish svartmylla [black soil], sandjord [sand soil], and starräng [sedge meadow], are not seen as names or name-like items. Such words as the Swedish träsk [moor] and ava [lake] in definite and indefinite forms are treated as potential microtoponyms. In the Pomeranian material, the lexemes Winterfeld [winter field] and Sommerfeld [summer field] seem to have functioned as names in some areas,¹⁴ but in other places were rather used as descriptive terms. I have excerpted all such items.

Microtoponyms are central to my analyses because these constitute a unique material attested in the large-scale cadastral maps and accompanying documentation for the first time. Thus, there is little authoritative tradition such as older records tied to these names, an important circumstance that makes their processing of much interest. The land surveyors who collected and wrote down the names are seen as directly responsible for adapting or not adapting them. The central role of the microtoponyms also means that the Swedish material containing only twenty-four potential non-Scandinavian possible candidates is excluded from many of the study’s analyses.¹⁵

Etymological Background, Secondary Name Formation, and Linguistic Adaptation

The study’s data is etymologically complex and requires adequate methods to capture and represent etymological uncertainties and ambiguities.¹⁶ Moreover, these uncertainties and ambiguities, such as secondary name formations, directly affect categorizations of place-name attestations as Scandinavian or non-Scandinavian or as linguistically adapted or not. Secondary name formation (i.e., coinage of new Swedish place-names containing older non-Scandinavian names, like the Swedish Luleåfjärden [Luleå bay] where the specific ultimately goes back to the Lule Sámi name of the Lule River) constitutes a methodological problem for this work. Should these names be seen as purely Swedish formations? Or is it of interest to represent the non-Scandinavian elements of the names in the data? If name usage in the seventeenth century is chosen as the point of departure, such names undoubtedly should be categorized as Swedish only. However, my interest in non-Scandinavian names justifies another approach (see Figure 2). The non-Scandinavian substrate in such names—Sámi, Finnish, German, or Slavic—is represented in relevant metadata in the underlying data sets, for example, language categorization. Thus, Luleåfjärden is categorized as Lule Sámi and Swedish.¹⁷ For this reason, it is important to distinguish between unambiguously non-Scandinavian names (MIN in Figure 2) and secondary name formations (included into MAX in Figure 2) when quantifying the data. For Pomeranian data, there are no differences between the MIN and MAX numbers due to lack of ambiguities. In Turku and Pori data, the distinction is relevant for all referent groups; the differences are either relatively small (377 against 398 attestations in settlements) or considerable (0 against 57 attestations in administrative units or 792 against 850 in microtoponyms). Secondary name formations are prevalent in the Västerbotten data, and thus the contrast between the two categorizations is high (forty-five against seventy-seven attestations in settlements, eighty-seven against one in administrative units, and twenty-four against zero in microtoponyms).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Distribution of non-Scandinavian attestations in total and by type of referent across the data sets. MAX includes secondary name formations; MIN does not include this group of attestations.

Any strictly binary distinctions with respect to linguistic adaptation would entail oversimplification in many cases. For instance, the field name Söderåker [Southern field] (MHA A 1b, 24) consists of Swedish lexemes and thus of course is of Swedish origin. However, Finnish microtoponyms, more specifically meadow names, such as Kuoppaniittu and Kaakkokorpi, built from kuoppa [pit] and niittu [meadow], kaakko [Southeast], and korpi [deep forest], respectively, are attested in the same map. This implies that the Swedish field name in question is likely a land surveyor’s translation or an independent rendering of the original Finnish name.¹⁸ To capture the ambiguities of this kind, I also chose a three-step categorization of placename attestations with respect to linguistic adaptation as either (1) nonadapted or (2) adapted (MIN in Figure 3), or (3) uncertainly adapted (MAX in Figure 3). Any examples of possible adaptation building on conjecture as the Söderåker example are included into the category uncertainly adapted. Thus, the Swedish microtoponyms like Söderåker or Karbowsheden (BI, 33, U420) built to the settlement name Karbow and hed [moor, forest] are categorized as adaptations to Swedish if the map includes Finnish or German microtoponymy.

In the following, I provide two numbers for adapted attestations, for example, 46/188 in the Swedish material, where 46 is the amount of certainly adapted non-Scandinavian name attestations, and 188 includes uncertainly adapted material and secondary name formations.

General Tendencies

Overall, there is a clear difference between the data sets with respect to linguistic adaptation of non-Scandinavian place-name material (see Figure 3).

In the Västerbotten data, all of the non-Scandinavian attestations are adapted (46/188 or 100 percent). This is consistent with the language contact situation described above and the substrate role of the Sámi and Finnish languages. As stated earlier, the majority of the adapted names are settlement names and names of administrative units such as Kukasjärvi and the secondary name formation Piteå socken, respectively.¹⁹ In most cases, these undergo phonological (44/78 attestations) and lexical adaptation (27/136). The included non-Scandinavian microtoponyms, for example, Luleåfjärden, represent secondary name formations and are not considered in further analyses below.

The Turku and Pori data occupy an intermediate position. If conjectures are included, 804 attestations or 62 percent of the total non-Scandinavian material (1,305 attestations) is adapted. The certainly adapted data constitutes 462 attestations or 36 percent of the relevant data (1,284 attestations),²⁰ whereas the 501 nonadapted attestations make up about 40 percent.²¹ Interestingly, settlement names of Finnish origin tend to be adapted, 54–60 percent or 214/240 of 398 attestations, mainly through epexegesis, that is, by attaching a Swedish generic indicating locality type to the Finnish name (e.g. Lähtisby, Potilaby, Vuojoki gård, and Herttolahemmanet containing the Swedish by [village], gård [estate], and hemman [homestead], respectively).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Distribution of adapted and nonadapted non-Scandinavian attestations across data sets in percentage. MIN includes certainly adapted non-Scandinavian attestations only. In MAX, uncertainly adapted material and secondary name formations are added.

Finally, the Pomeranian data in total include the least adaptations to Swedish overall, only 19 percent at the highest or 712/735 of 3,960 attestations in total. Over 80 percent, 3,225 attestations, are nonadapted.²² Of the two Pomeranian subsets, the non-Scandinavian name material in the maps barely shows any traces of the Swedish language. Only 7 percent of the attestations (49 of 720) underwent linguistic adaptation. The Pomeranian descriptions show a slightly different picture. There, 20–21 percent of the non-Scandinavian material, 663/686 of 3,240 place-name attestations, are adapted. As for the settlement names, only 14 percent, 270 of 1,932 attestations in the total of the Pomeranian data, are adapted, although the majority of the attested names are of Slavic origin and thus were as obscure to the Swedish land surveyors as the Finnish settlement names in Turku and Pori. The most prominent adaptation strategy for settlement names in Pomerania is explication, that is, specification of the locality type in the body of the text, for example, öfwer adelßgodzet Rub=kow (37, U497) [of the nobility estate Rubkow]. Although this strategy is similar to epexegesis used in Finland, the non-Scandinavian names themselves are not manipulated in Pomerania.

These results provide some general support to my hypothesis because adaptation rates in Västerbotten and Turku and Pori are somewhat higher than in the Pomeranian material. The Swedish data appear essentially different because all non-Scandinavian items undergo adaptation. Here, such underlying factors as type of language contact situation and place-name collection and representation practices play a more significant role compared with the perceived linguistic and cultural differences, because the name material in question does not include any items replicated contemporaneously. As for the Finnish and Pomeranian data sets, there is no difference in nature, only in degree. It is evident that non-Scandinavian place-names were left unchanged and underwent adaptation in both cases. In the data sets from both regions, it was likewise of interest to specify the type of locality for settlement names, although the proportions of adapted and nonadapted attestations between the data sets differ.

A similar picture unfolds if the adaptation rates of Finnish and Pomeranian microtoponyms are compared (see Figure 4). As mentioned earlier, microtoponyms constitute key material for my comparisons because in most cases these were collected, processed, and put on maps by the responsible land surveyors firsthand. In Turku and Pori, a maximum of 62 percent of the material (523 of 850 attestations) is adapted. However, if uncertain cases such as conjectures are omitted, the adapted attestations count up to only 29 percent (248 attestations). The nonadapted data comprise 38 percent or 327 attestations.²³ In the Pomeranian data set at large, the adapted microtoponyms count up to almost 40 percent (36–38 percent or 404/423 of 1,124 attestations), while more than 60 percent or 701 attestations are not adapted.²⁴ However, there is a significant difference between the map and the descriptions materials. Only 14 percent of the data (30 of 219 attestations) underwent adaptation in the maps, compared with over 40 percent in the descriptions (41–43 percent or 374/393 attestations). The nonadapted data constitute 86 percent (189 of 219 attestations) and 57 percent (512 of 905 attestations) of the total subsets, respectively.²⁵ Once again, the inherent uncertainties of the primary material, including the high proportion of conjectures in the Finnish data, complicates the analyses and the interpretation of the results.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Distribution of adapted and nonadapted non-Scandinavian microtoponyms across data sets, in percentage. MIN includes certainly adapted non-Scandinavian attestations only. In MAX, uncertainly adapted material and secondary name formations are added.

The Finnish and the Pomeranian microtoponym data sets provide sharper contrast with respect to adaptation strategies in use, see Table 1–4. Olof Mört and Magnus Bergman use translation in most cases when processing Finnish data, and phonological adaptation, explication, and independent rendering are also frequently used if uncertain cases are included into the analysis. On the other hand, if only certain adaptations are included, phonology clearly dominates the picture, and explication and translation follow in the second and the third places, respectively. The Pomeranian data, the two subsets included, shows a clear dominance of explication. The Swedish land surveyors in the region thus specify names’ referents in the body text in most cases. Apart from explication, morphological adaptation and translation are the most frequent strategies.

View this table:

• View inline
• View popup

Table 1

Adaptation Strategies Used in 850 Microtoponym Attestations in Turku and Pori

View this table:

• View inline
• View popup

Table 2

Adaptation Strategies Used in the Pomeranian Data Sets (Total, 1,124 Microtoponym Attestations)

If we consider these results against the background of the 1691 and 1697 directives asking land surveyors to Swedify non-Scandinavian names and appellatives, it appears that such Swedification was already under implementation in Finland and Pomerania in 1684–1695 and 1694, respectively, even before the directives were issued, although to a varying degree. The dominance of translation and explication clearly demonstrates the ambition to present and describe the names and their referents in the Swedish language. The adaptation rates and adaptation strategies, most importantly translation, used in Turku and Pori indicate that the land surveyors in Finland can be seen as more compliant with this place-name policy under development. The Finnish data set thus still occupies the intermediate position along the adaptation scale, between the Västerbotten and the Pomeranian data—if all the adapted data are considered. However, if certain adaptations are given priority, the Finnish and the Pomeranian data sets seem to be fairly similar with respect to adaptation rates in microtoponyms.

View this table:

• View inline
• View popup

Table 3

Adaptation Strategies Used in the Pomeranian Map Data (219 Microtoponym Attestations)

View this table:

• View inline
• View popup

Table 4

Adaptation Strategies Used in the Pomeranian Descriptions Data (905 Microtoponym Attestations)

The most popular adaptation strategies across the data sets—explication, translation, and independent rendering—are employed irrespective of the linguistic systems in contact. However, the preference for phonological adaptation in Turku and Pori and morphological adaptation in Pomerania is best explained if the systems in contact are considered. Typological dissimilarity and similarity between Finnish, Low and High German, and Swedish create favorable conditions for phonological or morphological adaptation. Being a Germanic language, Swedish lacks many of the phonological features of Finnish, which belongs to another language family. As a result, the Finnish diphthongs and consonant clusters are often simplified, for example, Paignekallenwore for Painokallionvuori possibly containing the Finnish nature name Painokallio to paino [press; weight] and kallio [rock] in genitive and vuori [mountain; hill]. Low and High varieties of German share vocabulary and morphosyntactic features with its northern neighbor. For this reason, the German place-name material is more easily integrated into the Swedish morphosyntax, for example, Mölendiket where the Swedish definite article -et is added (cf. Möhlendiek and Det kleine Feld where the German article das is substituted by the Swedish equivalent, Das or Dat kleine Feld).

Conclusion

This article has asked three research questions. The first one concerns methods or strategies to process non-Scandinavian place-names in a selection of Swedish large-scale cadastral maps from Västerbotten, Turku and Pori, and Pomerania. Two types of comparisons were carried out. The five data sets were compared in their entirety at first. In addition, this analysis was complemented by further investigations of microtoponym subsets from Finland and Pomerania. Microtoponyms constitute unique place-name data because most of them appear in the seventeenth-century maps and descriptions for the first time and therefore most likely were processed by land surveyors firsthand.

The analyses reveal that non-Scandinavian place-names can both undergo linguistic adaptation and remain unchanged in the data sets. In this respect, the Västerbotten material differs from the Finnish and the Pomeranian ones because it only contains adapted place-name attestations. There are no non-Scandinavian microtoponyms constituting primary formations in this data set. Any further comparisons between the Finnish and Pomeranian data posit a major challenge, because the Finnish material contains many uncertain cases of adaptation. If such cases are included into analyses, the Turku and Pori data set containing over 60 percent of adapted attestations occupies an intermediate position between the Västerbotten and the Pomeranian data. However, there are no major differences between adaptation rates in Finland and Pomerania if conjectures are omitted. This result becomes even more apparent if the microtoponyms are examined. There, over 60 percent at most and almost 30 percent at least of the Finnish material is adapted—to compare with adaptation rates in Pomeranian maps and descriptions, 14 percent and over 40 percent, respectively. These numbers persist if individual land surveyor behavior in Pomerania is examined.

This summary shows that the answer to the first research question is somewhat inconclusive due to complexities of the underlying primary materials, especially the Finnish maps. However, the focus on adaptation rates might be misleading. There is another parameter, frequencies of nonadapted attestations, that is free from uncertain adaptations because these are excluded. If these frequencies are compared, it becomes evident that Pomeranian data sets contain at least twice as many nonadapted attestations as the Turku and Pori data. In other words, nonadaptation rates offer a clearer picture of general processing strategies.

The second research question covers compliance with the 1691 and 1697 directives’ requirement to Swedify non-Scandinavian appellatives and names. As explained, Kammarkollegiet’s and the Land Survey’s codified policies most likely were in use long before they were formulated in writing. If Swedification is understood as linguistic adaptation in a broad sense—including explication and independent rendering—the study shows that most land surveyors did not entirely adopt this practice. Although examples of linguistic adaptation are attested in all three investigated areas, adaptation rates vary. Thus, the data sets can be placed along the Swedification or adaptation scale from the highest adaptation rate to the least in the following way: the Västerbotten data are followed by the Finnish data set, and land surveyors in Pomerania exhibit the lowest inclination to adapt names. It is important to once again emphasize that the directives of the late seventeenth century build on the actual practice of place-name processing established long before these were issued.

Finally, the third research question has to do with differences in strategies of linguistic adaptation. In Västerbotten, phonological and lexical adaptation strategies are the most frequent. In Finland and Pomerania, translation and explication dominate the picture. Land surveyors in Turku and Pori show preference for translation and phonological adaptation, whereas their colleagues in Pomerania make most use of explication and morphological adaptation. Linguistic and extralinguistic and social and cultural factors are in play here. Linguistic affinity, or, more precisely, linguistic features, of the contact languages explain preference for phonological or morphological adaptation. The phonotactics of the three Sámi languages (Lule Sámi, Pite Sámi, and Ume Sámi) and Finnish in the examined data sets appears to pose a major challenge for Swedish land surveyors, who often simplify diphthongs and consonant clusters. Phonology is likewise a point of departure for lexical adaptation in Västerbotten. In Pomerania, lexical and structural similarities between the German varieties and Swedish explain why High and Low German forms could be integrated relatively easily into the Swedish morphosyntax.

Linguistics alone cannot explain all the dissimilarities. Material characteristics of the sources have influenced the outcomes to a certain degree. As discussed, place-name attestations in the data sets come from maps and descriptions. The Pomeranian descriptions differ a lot from notarum explicatio of the Västerbotten and Finnish material, because these provide detailed information about the area, the settlements, and their economic resources, including features of the agricultural landscape. It is unsurprising that explication is the largest adaptation strategy employed here. What is surprising is the fact that this strategy is not used more often.

General characteristics of the language contact situations in question have also proven to be significant for land surveyors’ inclination to adapt names and for their choice of adaptation strategy. The Västerbotten material contains mostly non-Scandinavian settlement names replicated and attested centuries earlier, that is, elements of the Finnish and the Sámi languages represent substrate units in the Swedish seventeenth-century context, which explains the 100 percent adaptation rate. These settlement names transmitted orally and in writing were already established and thus had no need for further linguistic refurnishing. Although both Turku and Pori and Pomerania are characterized by living language contact, the language skills of the surveyors most likely differed. The four Pomeranian land surveyors most likely had good command of High and Low German, but it does not seem very likely that Olof Mört and Magnus Bergman, responsible for Turku and Pori, were skilled in Finnish. Moreover, the cultural underpinnings of the land surveyors’ language skills are important here. I argue that cultural affinity between Sweden and Pomerania or between the Swedish land surveyors and German informants explains why, for instance, the Slavic material in the German context does not trigger adaptation. These settlement names were perceived as German and thus left unchanged. On the contrary, many of the Finnish settlement names were provided with Swedish epexegetic additions. It is important to point out that these constitute newly established minor farms, and thus the epexegetic additions are most likely contemporaneous with the land-surveying procedures under consideration here. Overall, my initial hypothesis appears plausible, although uncertainties over the primary material interpretation affect its strength.

What does all of this mean in a larger seventeenth-century context? The concept of Swedification links the 1691 and 1697 directives to other seventeenth-century royal decrees and regulations (Nordin 2000, 78–80, 86–87, 134; also Petrulevich forthcoming a). Together, these constitute a developing Swedish language and place-name policy. Requirements to introduce Swedish as the language of state authorities existed in the provinces of Ingermanland, Narva, Livonia, and Estonia, in Narva’s case as early as 1615 (Nordin 2000, 285–86, 295–300).²⁷ The requirements to enforce practical implementation of Swedification, that is, of increasing linguistic and cultural homogeneity, can be seen as a rational attempt to introduce Swedish as a common standard language across the conglomerate state for effective administration. Moreover, it was an attempt to improve internal cohesion in the crown’s borders by creating a common supraidentity for diverse others, its multilingual and multicultural subjects.

It is generally held that the Swedish large-scale cadastral mapping project of the seventeenth century came about from an ambition to ensure further colonial expansion by seizing control of economic resources (Kain & Baigent 1992; Karsvall and Tollin 2010; Höglund 2013a). This article can add that these maps were also designed as ideologically charged devices aimed at creating and transmitting Swedified namescapes of the conglomerate state to generations of rulers to come. However, Swedification rates varied in practice depending on linguistic and extralinguistic factors, not least on linguistic and cultural affinity between the Swedish land surveyors and their local informants.