On the periodisation of early North Germanic (AAM/Post-Print)

Page 1 of 19 Accepted author manuscript, 25.11.2024, submitted to NOWELE 77.2 On the periodisation of early North Germanic Johan Schalin https://orcid.org/0000-0003-1546-2868 This article welcomes the recent proposal of Michael Schulte (2024) for a periodisation of early North Germanic, while pointing out problematic premises and criteria for the delimitation of the proposed subperiods. It is argued that in delimiting linguistic periods there is more to gain than lose for the clarity of scholarly argument if preference is given to linguistic rather than graphemic or socio-cultural criteria, even if it means that reconstructed, rather than ambiguously attested features, are given preference. The relevant chapter of a recent book by Nelson Goering (2023) is also reviewed to cast light on the syncope era. Elements and insights are proposed for a periodisation and labelling of early North Germanic suited for discussion on diachronic structural change. 1 Introduction A longstanding complication for studies in early (i.e. ancient and medieval) North Germanic1 is the lack of consensus on how to periodise and label historical stages and genealogical varieties of the language. Evidently, periodisation always involves arbitrary judgements, but in this case, divergence has also resulted from differing analyses of chronological and genealogical criteria. For example, it has been difficult to determine when North Germanic diverged from its NORTHWEST GERMANIC ancestor language. A related problem concerns the dialectal status of the Scandinavian runic inscriptions in the Elder Futhark (Nielsen 2000: 287–293, passim). Independent of this problem, there is no agreed-upon name in English for the oldest North Germanic offshoot (cf. German URNORDISCH). Further, there is no consensus on the point in time after which the North Germanic language(s) had evolved enough to justify the use of the temporal qualifier OLD, shared with classic OLD ICELANDIC and its closest late-medieval relatives. The qualifier OLD is often given to language stages beginning anytime between 700 and 1050. Scholars of adjacent disciplines appear comfortable with lumping together the early Viking Age and medieval language under a single term, such as OLD NORSE. Finally, concerning the in-between period, there are differing views on exactly when and in what sequence major structural changes transformed the oldest North Germanic language during an era stretching from ca. 475/550 to ca. 850 CE.2 More importantly, opinions differ on the weight given to purely linguistic arguments when postulating periodisations at the expense of other considerations, such as the graphemics of the runic alphabet or even language-external events, such as the beginning of the Viking Age, as defined by historians (Bandle 2002: 28). Recently, two authors have published research relevant to the question of periodisation. In this journal, Schulte (2024) published “Stationen der frühen nordgermanischen Sprachgeschichte: Zu einer neuen Periodisierung des Urnordischen”. In it, he proposes to divide the era corresponding to the use of the Elder Futhark in Scandinavia into four periods, based on a mix of phonological, morphological and “grapho-linguistic” criteria. Apart from this, Goering (2023) has published the book Prosody in Medieval English and Norse, which contains a chapter on “Vowel Loss in Runic Inscriptions” (2023: 164–188) that contributes to our phonological understanding of the transformational SYNCOPE ERA. “North Germanic” is not in this paper written in small caps because it is a well-known generic concept, neutral to periodisation and outside the consideration of labelling discussed here. 2 Hereafter “CE” will be omitted as redundant. 1 Page 2 of 19 Accepted author manuscript, 25.11.2024, submitted to NOWELE 77.2 After stating his aim of “trying to extract the main implications for foot structure” (p. 164), he examines and interprets the runic attestations, seeking support for a chronology of vowel reduction and foot structure. The aim of this article is to reflect upon whether the newly published ideas by Schulte and Goering provide an opportunity to seek convergence in the periodisation and labelling of early North Germanic. A case will be made for grouping the pre-documentary subperiods of North Germanic into linguistically motivated periods, so as to establish a terminology that is helpful for scholarly discourse on sequential structural language change, including on umlaut and syncope. In Section 2 a general overview is presented of significant distinctions and minimal common denominators of the existing proposals for periodisation. In Section 3 the arguments and conclusions advocated by Schulte (2024) and Goering (2023) are reviewed. Based on this, elements and insights are proposed in Section 4, which could form a basis for a consensual periodisation and labelling. 2 The problem and its background Figure 1 illustrates the periodisation and labelling extracted from the literature of the last fifty years, arranged by year of publication. Fifty years here represents the depth of research history deemed sufficient to track any possible recent convergence of opinion. Note that the handbooks that are listed tend to draw on well-known definitions from earlier research history (Widmark 2004 [2001]: 205; Karker 1996: 14–15, 29–31, displayed in conjunction with Jørgensen 2016: 79–83; O.E. Haugen 2012: 19; Nedoma 2010: 20–21; NSH IV 3: 10; Fulk 2018: 24). On top of these, Figure 1 also reflects original research into language history, where scholars pay particular attention to the periodisation problem (Antonsen 1994: 58–64; Grønvik 1998: 16–26; Nielsen 2000: 31–33, 98–99, 255, 287–293; Schulte 2018a: 13–15; Schalin 2018: 48–60; Schulte 2024: 53–61). 4 The works of Einar Haugen (1976: 123–127, 1982: 5) and the historical grammar of Voyles (1992: 35, 71, 103–104) have elements of both genres. Although periodisation is a recurrent topic in the voluminous handbook edited by Bandle et al. (2002: Preface VII–VIII, as well as chapters by Bandle 2002: 27–28; Birkmann 2002; Venås 2002: 32–36; Nielsen 2002: 615–618), a concise justification of the labelling used in the “Contents” section of it (Bandle et al. 2002: XIV–XV) seems to be lacking. 3 The periodisation table only occurs in the Preface to the series of volumes (Bull & Schulte et al. 2018: 10), but it would appear that the periodisation is reliant on the experts responsible for the respective chapters, namely Schulte & Williams (2018) and Schulte (2018b). 4 Tomas Riad (1992: 35–36) has Proto-Norse spanning to 1000 CE, but his periodisation has been omitted here because he states that it is “based on prosodic development only”, for the purpose of his exposition. Page 3 of 19 Accepted author manuscript, 25.11.2024, submitted to NOWELE 77.2 Figure 1. Overview of proposed schemes of periodisation and labelling of early North Germanic in the past 50 years. For paginated references see the first paragraph of Section 2. Page 4 of 19 Accepted author manuscript, 25.11.2024, submitted to NOWELE 77.2 Schulte invokes a purportedly existing consensus of two subperiods, called Frühurnordisch and Späturnordisch (2024: 49–53) as a backdrop to his proposed division of Urnordisch into four subperiods (Schulte 2024: 53–55). Figure 1, on the other hand, illustrates a remarkably incoherent picture suggesting there is little scholarly consensus on the number of periods, their temporal demarcation or their labels. The problems involved have been recognised and discussed from many angles, e.g. in Bandle et al. (2002, with chapters as referenced in the previous paragraph), as well as in Jørgensen (2016: 79–83), Nesse5 (2018: 35–39), Schalin (2018: 37–39, 48–60) and Schulte (2018a: 13– 15, with references). Yet, many handbooks are still happy with just summarising the research history of periodisation and with justifying their divisions of chapters instead of getting involved in argumentative original research on the subject or scrutinising the actual development of the language stages they aspire to describe. To create some order in the data given in Figure 1, the timeline will next be roughly divided into three slots by means of a non-linguistic terminology which is non-prejudicial towards the matter at hand: the Roman era, the post-Roman era, and the Viking era with the medieval era up until 1350. The existing proposals for periodisation, with significant distinctions and minimal common denominators compared, will be introduced in inverse chronological order. 2.1 The Viking era from the eighth/ninth century with the medieval era til 1350 The medieval era after 1100 falls strictly outside the scope of this paper, yet the period up until 1350/1375 is included to allow an examination of the temporal qualifier OLD, which concerns the medieval period just as much as the Viking Age. Only with the labels OLD ICELANDIC and OLD GUTNISH does the beginning of the usage of OLD coincide with the introduction of Latin script. In other combinations this qualifier typically applies earlier. For example, OLD SWEDISH [FORNSVENSKA] is claimed to encompass RUNIC SWEDISH [RUNSVENSKA], which begins already 800. The OLD DANISH [GAMMELDANSK] period has similarly been understood to start before attestations in Latin script, namely 1050/1100.6 NSH IV uses ELDER/EARLY OLD NORWEGIAN [ELDRE/TIDLIG GAMMALNORSK] even from 700.7 Projecting nationally-labelled languages this far back may be a convenient solution in handbooks on the history of codified national languages, but it is arguably anachronistic (Birkmann 2002: 691), does not serve clarity of scholarly discourse on the genealogy of these dialects, and does not – for good reasons – ever occur in sources with a pan-Scandinavian perspective. The solution, moreover, appeals to the constructed “Viking” narratives inherited from the national-romantic period. Nothing in the better-known dialect geography from the middle ages suggests that dialect splits in the previous stage coincided with modern state borders (Bandle 2011: 18–23).8 Cf. note 3 above. The earlier period of OLD DANISH [GAMMELDANSK] is confusingly called EARLY MIDDLE DANISH, while the preceding period (800–1050/1100) may be called either RUNEDANSK or (equally confusingly) OLDDANSK (Jørgensen 2016: 80–82), the latter translated in Figure 1 as ANCIENT DANISH. 7 The foreword (Bull & Schulte & al.: 10) uses ELDRE GAMMALNORSK, while the relevant chapter (Schulte 2018b: 123, 135) uses TIDLIG GAMMELNORSK. Nesse (2018: 38) uses both ELDRE and TIDLIG. 8 One reviewer interprets a recent study by Dagfinn Skre (2022) to imply support for an idea that the “location of ruler sites in Scandinavia between 300 and 800 shows three or four focus areas, that coincide with later Denmark, Sweden (Svealand), Oslofjord and the Norwegian west coast” and that the “later kingdoms are no coincidence”. The reviewer further reminds that language “as a social phenomenon tends to be grouped along the line of societal organisations = (pre-)states”. None of these ideas are, however, to be found as such in the source, where e.g. Denmark counts 7 Viking Age ruler sites scattered over several regions of Denmark, two being in present day Swedish Skåne (Skree 2022: 118–119). On dialectal innovations it is legitimate to generalise that innovations radiate from prestigious centres, but in the Viking Age there were no national borders where the spread of these 5 6 Page 5 of 19 Accepted author manuscript, 25.11.2024, submitted to NOWELE 77.2 Whenever national genetic designators are avoided, there is no agreement about which alternative to use: SCANDINAVIAN, NORSE or NORDIC. In such cases the use of OLD is backdated to at least 1050, at which point E. I. Haugen defines the beginning of OLD SCANDINAVIAN. However, more often OLD is backdated even farther, at most to 700, which is when O.E. Haugen places the beginning of ELDRE NORRØNT [ELDER OLD NORSE]. In a wide sociolinguistic sense, these languages were all considered dialects the same language, demonstrably referred to as “Dǫnsk tunga” [Danish Tongue]. Nonetheless, the North-Atlantic or Old Norwegian varieties spoken in Iceland, western Norway, as well as the Faroe, Orkney and Shetland Islands were much closer to each other than to the others, and the use of NORRǾNT MÁL, corresponding to the modern English term OLD NORSE, has strong connotations to this close dialect group (Schulte 2018b: 125). 9 This is reflected in a preference among many scholars, especially from Denmark or Sweden, to use NORDIC or SCANDINAVIAN for any referent with wider scope. Such practice, which limits the use of OLD NORSE geographically and temporally to the post-Viking Age LATE OLD WEST NORDIC of the Icelandic literature, has the advantage of resolving severe terminological ambiguities and contradictions involved, to wit: OLD NORSE cannot at the same time be both a thirteenth century descendant of a Viking Age OLD WEST NORDIC and the ancestor of CLASSIC OLD SWEDISH and EARLY MIDDLE DANISH, which are contemporaneous with it.10 Early twentieth century research used to put a limit between two overarching eras, URNORDISCH and ALTNORDISCH, at the year 800, which approximately coincides with the beginning of the Viking Age (Venås 2002: 33; Birkmann 2002: 692). Danish and Swedish handbooks maintain this tradition. Yet the language of the Viking Age is not as static (or uniform) as often perceived, and it is doubtful whether it delimits a meaningful unit in linguistic periodisation. Accordingly, the year 800 has been nudged in both directions. Some, including O.E. Haugen (2012: 19) and Schulte (2024), would like to set it around 700 (or 700/750) to make it coincide with the heyday of the orthographic transformation of the runic script, while others (Grønvik 1987:167–189, 1998: 16–26, cf. Myrvoll 2022: 185) would prefer to move it further into the Viking Age to 850 to mark the end of the phonologically defined SYNCOPE ERA. The more fine-tuned periodisation by Schalin (2018: 55) delimits a period 600/750–825/900, which is intended to fall between the earlier waves of syncope and the last vowel shortening, albeit the absolute dates are expressed unnecessary widely. An outlier is Einar Haugen (1976: 135, 1982: 5; cf. the figure in Nedoma 2010: 20) who proposed to lump together the Viking Age era (ending in 1050) with the previous instead of the following era, thereby introducing the concept of COMMON SCANDINAVIAN for 550–1050. The use of COMMON is clearly problematic, since the term may be misunderstood to mean a common ancestor, a “protolanguage”, and the last centuries of this period encompass the splitting up between one western and an unknown number of eastern speech varieties (Venås 2002: 32; Bandle 2011: 18–23). This is the case although the contemporaneous speakers evidently identified these dialects as varieties of the same language. It is very important for diachronic research to establish an analytical periodisation that in retrospect reflects the actual branching of the language. A trade-off against a better sociolinguistic description would in this case be quite unhelpful, as evident from the discussion above referring to note 10. innovations might have stopped. Diversity in Sweden, as evident in the different outcomes of umlaut in Gutnish and Elvdalian, would have been manifest already in the Viking Age (cf. the illustration of this period in Figure 2). 9 The term is well-established, reflected by the fact that it enjoys an ISO 639-2 and 639-3 identifier “non” for a historical language (https://iso639-3.sil.org/code/non), albeit without further temporal or geographic specifications publicly available. 10 Examples of this widely ignored paradox include the classification in the Glottolog 5.0 ‘Spoken L1 Language: Old Norse’ (2024) and of Wikipedia ‘Category:Old Swedish language’ (2024). Page 6 of 19 2.2 Accepted author manuscript, 25.11.2024, submitted to NOWELE 77.2 The post-Roman era; from the fall of Rome to the eighth/ninth century The preceding period beginning around 475/550 up until the eighth/ninth century is marked by very rapid and transformative language change, which is only partially traceable in the runic data. A quite frequent practice is to denote this as a late version of URNORDISCH, by the terminology of Grønvik (1998: 16–26) NACHURNORDISCH, by Rischel (2008: 192) LATE ANCIENT NORDIC, by Schulte (2018a: 13–15) SPÄTURNORDISCH, and by O.E.Haugen (2012) YNGRE URNORDISK [YOUNGER PRIMORDIAL NORDIC]. Sometimes, and more often so in Danish and Swedish manuals, the period, often delimited up until 800, is lumped together with the previous one without adding a qualifier of lateness at all, as done by Karker (1996: 14–15, 29–31) using FÆLLESNORDISK [COMMON NORDIC], by Widmark (2004 [2001]: 205) using URNORDISKA, and by Fulk (2018: 24) using RUNIC (PROTO-NORSE). Conversely, Einar Haugen (1976: 135, 1982: 5) disconnected this period from the previous one and prolonged it to 1050, as already mentioned in subsection 2.1. More fine-tuned linguistic subperiods of the post-Roman era have been proposed by Grønvik (1987:167–189, 1998: 16–26: NACHURNORDISCH stages I, IIa, IIb and III) and Schalin (2018: 48–60: POST-PROTO SCANDINAVIAN and TRANSITIONAL SCANDINAVIAN). Both attempted to create temporal qualifiers that could facilitate diachronic discussion on the stepwise progressing process of umlaut and vowel loss. It is obvious from these examples, as well as from Figure 1, that the alternative genetic designators SCANDINAVIAN, NORSE or NORDIC are used quite interchangeably, albeit each with its own connotations and proponents. I suggest using NORDIC for the later language stages, during which the geographically non-Scandinavian areas outside Denmark, Gotland and the Scandinavian peninsula were colonised, while SCANDINAVIAN can be reserved for the language stages contemporaneous with early runic inscriptions and spoken in what today is understood by SCANDINAVIA.11 This would also help to keep the two overarching periods intuitively apart. The designator NORSE would be disambiguated to hold exclusively the meaning which it carries in the most concrete contexts, i.e. as used when indicating provenance of lexical material of the LATE OLD WEST NORDIC literary language, which in its richness deserves identification with this well-branded label. 2.3 The Roman era: from the appearance of runes in Scandinavia to the fall of Rome In the Roman era phonological change was slower. For some time after East Germanic had diverged from PROTO-GERMANIC, common NORTHWEST GERMANIC innovations occurred before this later unity in turn broke up into West Germanic and North Germanic (Nielsen 2000: 287–293; Schulte 2018a: 18– 20). The idea that the early runic language in effect was NORTHWEST GERMANIC had famously been proposed by Kuhn (1955), who wrote in a period somewhat earlier than the research covered in this paper. In Figure 1, Antonsen (1994: 58–63, cf. 1975) is instead quoted as a younger and often cited representative of this opinion. Further supporters of variations of this position may be found in Fulk (2018: 16, with references). Nielsen (2000: 57–65; cf. Schulte 2024 66–68) shows how research that appeared after Kuhn struggled to determine when North Germanic had separated from North-Sea Germanic and the other West Germanic languages. Reluctance towards equating early runic with NORTHWEST GERMANIC language reveals itself in two ways: labelling and periodisation. Some of the German and Scandinavian authors (including authors of In Europe, the modern connotations of SCANDINAVIA and SCANDINAVIAN are clearly inclusive of Denmark and Gotland. When it is necessary to specify the exclusion of these territories, the term SCANDINAVIAN PENINSULA is used. Conversely, Icelanders and Finns do not today self-identify with the qualifier SCANDINAVIAN but use equivalents of NORDIC (or NORDEN) for that purpose. 11 Page 7 of 19 Accepted author manuscript, 25.11.2024, submitted to NOWELE 77.2 many handbooks), who just continue to use URNORDISCH or URNORDISK(A), do not take a position on the relation of this language to NORTHWEST GERMANIC. Differently, Einar Haugen (1976: 123–127, cf. 1982: 5) accepts Antonsen’s view on early Runic, calling it “only mildly Sc[andinavian]” (1976: 127), but nevertheless uses PROTO-SCANDINAVIAN for labelling a contemporaneous reconstructed language. A mirror image of this approach is represented by Dirk Boutkan (1995: 31–35), who lists features showing that the early runic language “was not the parent stage of Ingveonic but may be considered a proto-stage of N[orth] G[er]m[anic]”. Still, Boutkan himself settles for OLD RUNIC, justifying his decision by saying that it avoids using NORTHWEST GERMANIC. Hans Frede Nielsen (2000) goes to great effort to place the early runic language in a dialectal (Northwest) Germanic context. Nielsen (2000: 31–33) justifies his use of the name EARLY RUNIC (for 200–500) “in order to avoid terminology which prejudices the provenance issue”.12 Nonetheless the conclusion arrived at by Nielsen (2000: 293–296) is that “the relative proximity of Early Runic to North Germanic (Norse)” is characterised “as being of the first degree” and “that to North-Sea Germanic (Old English) of the second degree” and, finally, “that of Early Runic to Old High German of the third degree” (2000: 375). Nielsen (2000: 100–103; 294; 370) found that the EARLY RUNIC vowel inventory in unstressed syllables had developed so that only the contrasts that existed in North Germanic vowels, as opposed to those of North-Sea Germanic, could be derived from it. Nielsen (2000: 290–293) still sees this as a very modest first step in the development of North Germanic and does not want to go all the way towards readjusting the labelling practice. This becomes evident when, in his chapter in Bandle et al. (Nielsen 2002: 618), he states his reservations about the term ANCIENT NORDIC, otherwise favoured by the monograph, by expressing that the period partly “antedates the inception of specifically Nordic features in the sixth century”. By means of reconstruction, other relevant criteria may be found than those most discussed in the context of runic inscriptions. An important example is the early North Germanic final obstruent devoicing and the sound changes that must have preceded it (Ralph 1980: 5–9). Also, the North Germanic rhotacism of /z/ proceeds in different stages and has different effects in some phonological environments compared to the West Germanic ones (e.g. English hoard, German Hort, Old Icelandic hodd), which shows that it did not happen in NORTHWEST GERMANIC (Schalin 2018: 140–147). Compared to the proliferation of proposals to classify early runic language one way or the other, there are fewer suggestions to divide the Roman era on a timeline between NORTHWEST GERMANIC and early North Germanic. Where Voyles (1992) and Schalin (2018) do so, this is not done in a runic context, but in that of a reconstructed ancestor of the Scandinavian languages. Precisely how closely EARLY RUNIC can be equated with such an ancestor is far from self-evident (Haugen 1976: 124). An important piece of evidence to set a terminus ante quem for the separation of West and North Germanic is the inscription kaba on the Frienstedt Comb, which dates to the second half of the third century (Schmidt et al. 2012: 131). The four runes undoubtedly represent the word m. sg. nom. ka(m)ba ‘comb’ and the case marker -a < PGmc *-az represents the oldest attested West Germanic dialect feature that sets the language apart from North (and East) Germanic (Schmidt et al. 2012: 139–158; Schulte 2018a: 37). This find was not known to Antonsen, but he might not have been too bothered by it since he argued (Antonsen 1994: 58–63, cf. 1975) that the archaic runic language of even later Scandinavian inscriptions was better called by the same name as its own ancestor – “Northwest Germanic” – even if some West-Germanic dialects had already split off from it.13 In my view, a periodisation should – to Nielsen (2000: 38) further explains in nt 21: “Since an important aim of my project is to investigate in depth whether Early Runic can, in fact, be the ancestor of both Old Norse and (some of) the West Germanic languages, a linguistically neutral term is obviously called for”. 13 The analogy invoked by Antonsen (1994: 62) was that Dutch did not cease to be Dutch just because Afrikaans became a language in its own right. This analogy is superficially ingenious, yet in its context unfortunate, to wit: of course, a split of Dutch would have been posited in a prehistoric context, and why not in modern times as well, had it not been for the political role that codified language have acquired in nation states. 12 Page 8 of 19 Accepted author manuscript, 25.11.2024, submitted to NOWELE 77.2 keep things clear – recognise a split by renaming both descendants. Even if the innovations in one of the descendants do not happen to reveal themselves in scarce attestations it does not mean that they did not exist: all languages are constantly changing. Finally, for the Roman area, the terms COMMON and PROTO-, which elsewhere in diachronic linguistic discourse generally indicate hypothetical common ancestry, are best avoided for this language stage. An imposing example to this effect, an elephant in the living room so to say, is the label PROTO-NORSE. While not prominently appearing in Figure 1 nor in too many linguistic works, it is frequently used elsewhere. Despite persistent statements of disfavour (e.g. Antonsen 1994: 58; Nielsen 2000: 32; Schulte 2018a: 13–14; Goering 2023: 165) it has clearly gained a foothold among scholars from adjacent disciplines, such as philology, runology, archaeology, Viking studies and Old Icelandic literature.14 The main vector of criticism against it is the fact that PROTO- (and to some extent COMMON) in historical linguistics has the status of a technical term, which refers to unattested and hypothetical ancestor languages of language families reconstructed by means of the HISTORICAL-COMPARATIVE METHOD. Here the problem is twofold: runologists are the first to recall that early North Germanic is neither unattested nor hypothetical, an argument that I will argue to be problematic in subsection 3.1. Rather, in my view, comparativists should sound alarm, since the level of reconstruction reached by comparing the adequately documented medieval languages would not be the Roman Age language that is meant here, but a later language stage in the process of leaving the early period(s) of umlaut and vowel reduction (Nielsen 2000: 59). In this sense so called PROTO-NORSE is effectively a pre‑proto language achievable at best through internal reconstruction and even better through comparison with other Germanic languages. Also, the use of NORSE in this context is unashamedly Norway-centric and therefore arguably anachronistic. 3 The recent contribution to periodisation by Michael Schulte Michael Schulte (2024: 49, 54) presents a new periodisation of the language of the older runic inscriptions (until 700/750) by dividing the period of the preceding seven centuries into four subperiods: 1. Archaic, Pre-Classic Urnordisch (ca. 1/50–250/300) 2. Classic Urnordisch (ca. 250/300–400/450) 3. Late Classic or Post-Classic Urnordisch on the threshold of the syncope era (ca. 400/450– 550/600) 4. Late Urnordisch of the transitional and syncope era (ca. 550/600–700/750) He thus suggests we accept a period of transitional inscriptions, as proposed by Barnes (1998), and treat it as a linguistic era and, further, split it into two, which yields a POST-CLASSIC and a LATE URNORDISCH subperiod – the two being differentiated by a change of the phoneme-grapheme correspondence of the Jāra rune from /j/ to /a/. At the other, early end of the timeline, Schulte utilises some rune-typological observations to support the delimitation of his ARCHAIC, PRE-CLASSIC subperiod, while referencing Nielsen (2000: 154, 282) for a linguistic criterion. Schulte deserves credit for initiating a discussion aimed at defining a more fine-tuned periodisation of early North Germanic. It is long overdue. He is also well placed to herald it, given his expertise in historical phonology and runology. He is not the first one to try, though. Ottar Grønvik (1987:167–189, 1998: 16–26), for example, paved the way for developing this topic by proposing an era of NACHURNORDISCH between 475 and 850 CE with further subdivisions into several developmental A symptom of the successful branding of this misnomer is the choice of Syrett (1994) to have it in the title of his dissertation, just to tell the reader that he will use “urnordisch” instead of “proto-Norse” in the text itself (1994: 29). 14 Page 9 of 19 Accepted author manuscript, 25.11.2024, submitted to NOWELE 77.2 stages labelled by numbers. Schulte also deserves credit for delimiting periods and subperiods in terms of chronological criteria rather than absolute dates. Methodologically this is better justified. Delimiting subperiods is not an end in itself. Schulte (2024: 61) recognises this when suggesting that a future use of a new periodisation could appear in handbooks and other historical linguistic publications. Schulte does not, however, propose another easily conceivable and important use for a better periodisation, such as the fact that a more precise and consensual terminology would benefit clarity of scholarly debate and analysis. This purpose corresponds to a need in historical phonology and morphology to represent reconstructed forms of words from all stages of language development. This includes short-lived forms from brief stages of development, which are of key importance to analyse structural change during the umlaut era, on which research is still ongoing (Schulte 1998; Kiparsky 2009; Myrvoll 2012; Schalin 2021, 2024; Goering 2023: 184–186). This need is clearly demonstrated in how I elsewhere (Schalin 2017: 192–194; cf. Figure 1 above and Schalin 2018: 48–60), in the absence of a conventional fine-tuned periodisation, had to define short subperiods to be able to label such forms when tackling the research questions.15 3.1 Appraising Schulte (2024) When Schulte (2024: 50) introduces his argument, he spells out his periodisation methodology by arguing for an approach, where each sub-period “must be distinguished by at least one reliable differentia specifica”. These are “primarily phonological and morphological criteria, but they also include the phoneme-grapheme relations of individual runes and their typological forms, in other words grapho-linguistic criteria”. Where Schulte (2024: 52) discusses the issue, he emphasises that combining grapho-linguistic with purely linguistic criteria stems from necessity, or to translate his own words: “Methodological rigour has to give way to the compromise of a mixed approach because of the meagre state of attestation”. To a certain point I agree with this statement, but compromising methodological rigour is a good servant but a bad master, and combining criteria must not be allowed to jeopardize the integrity of the argument. The context where Schulte (2024: 52–53) invokes the meagre state of attestation to justify the necessity of introducing grapho-linguistic criteria is when he argues against Myrvoll (2022: 185), while making his case for the termination of URNORDISCH at 700/750 instead of ca. 850 when the SYNCOPE ERA came to an end. The context is not well chosen since precisely here the evidence is not meagre at all but quite unambiguous, which Schulte in fact acknowledges. Whatever their varied dialectal status may be, the chronologically relevant inscriptions of Sölvesborg (Bl 5), Helnæs (Fyn 8), Rök (Ög 136), Kälvesten (Ög 8) and Hangvar (G310) agree on the retention of -u in the accusative of the light stem sunu after the first waves of syncope. Moreover, the nominative sunuʀ is attested in Gursten (Sm 144) and mɑkuʀ in Sparlösa (Vg 119) (Birkmann 1995: 178; Goering 2023: 176–177).16 For example, from which language stage is Finnic kari ‘skerry’ borrowed (cf. Old Swedish neut. skær), presuming that the loan originally had a form [skari] after early apocope but before front umlaut? Using terms such as “Early Late” Urnordisch or “Middle Post-Classical” Urnordisch would not make things any clearer, even if these sorts of clumsy qualifiers would follow from some of the discussed periodisations. 16 Goering cites sitiʀ in the Rök inscription here, in line with most scholars. An alternative should, however, be recognised, that it rather belongs to his examples of vocalised glides, such as fiɑru and garuʀ (kɑruʀ) in Rök, as well as sækiʀ (sɑkiʀ) in Oklunda (Ög N288) (Goering 2023: 179–180). This is because the present (and infinitive) stem, which surfaces in forms with back-vocalic endings, is sitj-. Commonly, the West Germanic sound law *Cji > *Ci is assumed to have become effective before the separation of North Germanic (Voyles 1992: 74–124; Ringe 2006: 129–130), but this is not required by the North Germanic data. In Schalin (2021: 27–28) it is pointed out that Old Swedish front umlaut in 3.pers.sg. *swær points to a glide in the syncopated ending, since it contrasts with 3.pers.sg. far < *fariʀ, which could be taken to indicate that the sound law *Cji > *Ci never applied. Analogy 15 Page 10 of 19 Accepted author manuscript, 25.11.2024, submitted to NOWELE 77.2 Even if Schulte in important aspects relies on Barnes (1998) – referencing him eleven times in three different chapters – he nonetheless bypasses without comment a main premise of that paper, which seems especially relevant to his “mixed approach”. Namely, Barnes (1998: 460) sets as a condition for plotting “rune-typological and linguistic change on the same continuum” a requirement to “cease to confuse runic typology with linguistic typology and both with archaeological dating”. With regard to the “linguistic changes that divide Primitive Norse from Old Norse and the graphic changes that come between the older and the younger futhark”, Barnes (1998: 449) states that “only one of the two sets of changes should be chosen to provide the criteria by which an inscription is assigned transitional status”, before going on to note that the “choice will be a matter for individual scholars, reflecting their particular aims and interests.” Further Barnes (1998: 454) does not see how transition “can be defined in terms of runic form, language and time without the twin dangers of circularity of argument and confusion.” Consequently, Barnes (1998) discusses the temporal qualifier TRANSITIONAL only in conjunction with “inscriptions” rather than with language stages. 17 Here the reader would have benefitted from a clarification by Schulte.18 In my view, when plotting “rune-typological and linguistic change on the same continuum” (Barnes 1998: 460), two conflicting considerations should be kept in mind. First, in line with methodological rigour, it is necessary to keep apart linguistic testimony of early runic writing from reconstructed idealised variation-free early North Germanic; also, it is well to keep in mind the fact that neither of these two are necessarily identical to the varieties that were in fact spoken.19 As for these two sources of knowledge, we have a fair amount of understanding from work on many language families of the homogenisation and risk of anachronous distortions created by the comparative method and internal reconstruction. Conversely, we have few reliable means to recover the factors that influenced rune carvers. Because of the scarcity of attestations, we can hardly categorise reliably the cultural context and the writing conventions used in each inscription. On top of this, we can hardly relate the runic typology to the variation of spoken dialects and sociolects, not to mention other Germanic standards perhaps known in the area. Some of the variation in the early inscriptions is likely to reflect geographical and dialectal variation that has been eliminated in the radical and remarkably uniform development of North Germanic in the post-Roman and Viking Age eras (Schalin 2018: 124). For example, Syrett (1994: 32) contrasts the root vowels in Tjurkö (KJ 136) wurte, kurne against those in Gallehus (KJ 43) horna and By (KJ 71) orte. Further, three different patterns of Sievers’ law emerge depending on whether: 1) syllabification is reconstructed regularly, based on syncopated forms, 2) inferred from Sámi loanwords with an expected syllable missing in heavy stems (Luobbal Sámmol Sámmol 2020: 9), or 3) established in runic inscriptions with one syllable too many in light stems, examples being many personal names (talijo Talijō DR 206; harija Harija KJ 85/Sö 32; Þrawijan Þrawijan KJ 61) and the appellative gudija in may of course be invoked in many ways to explain this data, one way being that the glide was restored throughout the paradigm in early North Germanic due to levelling. 17 Similarly, but inversely, Schalin (2018: 52) defines PROTO-SCANDINAVIAN independently of runic language “in order to maintain methodological rigour and secure the autonomy of reconstructive methodology”. 18 The need for clarification is of course in no way satisfied by the published e-mail from Barnes to Schulte (2024: 72), where Barnes inter alia states the obvious regarding Schulte’s approach, namely that given the different aims, it is “unsurprising that we arrive at different results” and concludes conditionally: “If one accepts that a clear quadripartite division of the older fuþark can be based on limited and sometimes disputed evidence, then I see no objection to what you propose.” 19 Haugen (1976: 124) suggests that “the term P[roto-]Sc[andinavian] should be reserved for the reconstructed late N[orth] G[er]m[ani]c (or N[orth]W[est] G[er]m[ani]c) which is the ancestor of the later Sc[andinavian] languages. It is not at all certain that these two – P[roto-]Sc[andinavian] and Runic – are identical”. Further, Antonsen (1994: 59) states “the basic fact that not all the inscriptions in the older fuþark are written in the same language”. Page 11 of 19 Accepted author manuscript, 25.11.2024, submitted to NOWELE 77.2 the Nordhuglo (N KJ 65) inscription (Syrett 1994: 185–186; Fairfax 2015: 166).20 Whatever the reasons for these variations, they are not recoverable by reconstruction. In all, it is likely that dialectal variation and perhaps graphemic-phonemic mismatches are accountable for some of the variation in the linguistic testimony of inscriptions, but methodologically this cannot be adequately examined (Syrett 1994: 29– 35). For these reasons, when discussing the genealogy of early Germanic and the descendance of lexical material and grammatical structure from PROTO-GERMANIC to its medieval North Germanic descendants, we must primarily apply the tools of historical linguistics to reconstruct an idealisation of a spoken intermediate primordial North Germanic language stage. For this line of research an adequate periodisation and labelling of structurally defined language stages is a vital analytical instrument, and not one whose purpose is merely to organise data or to serve as a descriptive category in handbooks or as a didactic tool. For diachronic linguistic research, a periodisation aspiring to define language stages by invoking rune-typological criteria is unhelpful. Now, having elaborated on methodological rigour we shall turn to a conflicting, equally important consideration. More in agreement with Schulte’s motivations, I do believe that strictly maintaining a firewall between the two ways of acquiring knowledge – the comparativist track and the runological track – is a luxury that we cannot afford, even if maintaining such a firewall ideally would be methodologically desirable. The two tracks cannot proceed in full isolation if we aspire to fill in some severe knowledge gaps that persist in both. As with any multidisciplinary endeavour the account must, however, be clear about which methodology any particular statement is based on and be transparent about a hierarchy of relevance and reliability of evidence when reconciling results. In this vein, inscriptions have been successfully used, and indeed should continue be used with due diligence, as relevant but problematic testimony to the language situation in any given period and area. The problems with the periodisation proposed by Schulte are in my view twofold: first, on a conceptual level, a periodisation based on runic typology cannot aspire to function as a periodisation of a language, and the labels used (PRE-CLASSIC, CLASSIC, POST-CLASSIC) would indicate that this was perhaps not even a main aspiration of the author: where there are good grounds to identify a subperiod of landmark inscriptions from ca. 250/300–400/450 AD as prototypically classic, there are hardly any grounds to assign the contemporaneous stage of spoken language such classic status. Second, Schulte (2024: 50, 52-53, 73) mixes into his overall concept of “grapho-linguistic” criteria two entirely divergent sets of criteria, one of which is admissible testimony of language change (phoneme-grapheme relations) and one which is not (rune-typological criteria). Consequently, a valid delimitation of CLASSIC from POSTCLASSIC URNORDISCH at 400/450 AD is less solidly established, in that the change in the shape of the Jāra rune is irrelevant to language change.21 Schulte’s use of another criterion to mark the start of the POST-CLASSIC subperiod, i.e. incipient syncope (Schulte 2024: 51) brings forward the beginning of the SYNCOPE ERA while including two inscriptions with apparent third syllable vowel loss. In this regard it is a matter of definition whether to interpret the form fahi fāhī (< *faihiju) in the Noleby inscription (KJ 67) and hroʀeʀ hrōʀēʀ/hrōʀīʀ in the By inscription (KJ 71) as late cases of the older Germanic Auslautgesetze, or as early cases of trigger loss belonging to the incipient transformational (umlaut-syncope) period of early North Germanic. 22 A structurally motivated criterion to distinguish between the two would be to define the beginning of the latter by the phonemicisation of the first umlaut vowels in the vowels inventory. This may first have been triggered by loss of third syllable /i/ in words with a syllable structure equivalent to *anudiʀ > For a commonly occurring second part of names -warijaz, Bammesberger (2008) proposes it to be a heavy stem. 21 The change in its phonemic correspondence comes later and is used by Schulte to delimit the POST-CLASSIC subperiod from LATE URNORDISCH. 22 The form writu wrītu in the Eikeland inscription (KJ 47) is not equivalent as the retained -u would have resisted second syllable apocope. 20 Page 12 of 19 Accepted author manuscript, 25.11.2024, submitted to NOWELE 77.2 *ænudʀ ‘ducks’, namely in an environment where none of the constraints holding back second syllable vowel loss would have been active (Kiparsky 2009: 19–26; cf. Goering 2023: 167–170). This loss can hardly be earlier than loss of -a- in the word erilaʀ/irilaʀ, which occurs unsyncopated alongside hroʀeʀ in By.23 Thus understood, the umlaut-relevant SYNCOPE ERA is likely to have commenced later than the By inscription. Nonetheless, it could turn out motivated to delimit a fifth century subperiod, in which some changes may be reconstructed that are precursory to and prerequisites for explaining the subsequent colouring umlauts. Schulte (2024: 55–57) presents mixed arguments for a separate PRE-CLASSIC period. As far as the language rather than the rune shapes are concerned the nominatives of the masculine n‑stems in -ō are undoubtedly an admissible criterion for early age, already identified by Nielsen (2000: 154, 282). Strictly, its use for dating purposes is, however, limited by the fact that it cannot be contrasted with anything following it except its absence in the data, which might be coincidental. The later nom.sg. endings of masculine n‑stems do not serve the purpose since they may have an ancestor (contemporaneous with the ending -ō) represented by the -a on the Vimose comb and on the shield handle 1 from Illerup Ådal. Schulte (2024: 56–57; cf. 2018a: 27–32) discusses this parallel ending exhaustively without conclusively establishing its origin or descendant. This criterion for separating a PRE-CLASSIC era raises a related and most intriguing issue, namely whether the language this early can be identified as North Germanic at all. To justify the subperiod as an URNORDISCH one, Schulte (2024: 66–67) invokes a recent find from Svingerud/Hole, with the reading faḥido fāhidō seemingly testifying to early monophthongisation in a phonological context typical of North Germanic. 24 Schulte accounts well for the fact that in other inscriptions the same monographic spelling has been dated later. In discussing this reading, he could also have added that the “dating of the part identified as Hole 3 is uncertain and find circumstances are insecure” [..] “Further archaeological analysis is needed of the detectable structures of the grave field and their potential links” (Zilmer & Vasshus 2023: 285). 25 According to personal communication from Krister Vasshus (8 October 2024) such analysis may be forthcoming and I thus leave it to future research to judge what dating is most likely for this testimony, given that it seems very early as a witness for monophthongisation. At present, it is premature to fix a dating earlier than that for kaba on the Frienstedt Comb (see subsection 2.3) and, thus, to bring forward the end of Northwest Germanic which can otherwise be more plausibly dated as late as the second or third century. In any event Schulte is likely right to assume that a change such as this monophthongisation has not spread to all parts of the speech area at once. At the other late end of the timeline, Schulte’s delimitation of LATE URNORDISCH in his narrow sense relies completely on the new phoneme-grapheme relation of the Jāra rune to any of the oral low vowels. It is a legitimate proposal based on an admissible criterion, but an atomistic one, not referring to a major structural change in the language. It would better serve the purpose to choose a change with more structural implications, such as the first umlaut-relevant syncope or some subsequent phonological development unleashing a new wave of trigger loss (cf. Kiparsky 2009: 19–26) entailing further diversification of the vowel inventory. An earlier claim by Schulte (1998: 229f), that hroʀeʀ in By would testify to a merger of unstressed /ī/ and /ē/, which in turn would have caused a very early palatal umlaut, seems a too far-reaching conclusion to be based on one rune, considering the uncertainties of the graphemic-phonemic fit (Syrett 1994: 35) and the later representation of a vowel of the same origin by an i-rune in hᴀeru-wulafiʀ hʲeru-wulᵃfīʀ < *heru-wulfijaz in the Istaby inscription (DR 359, KJ 98). 24 On the face of it, it reminds one of many infamous cases where we are in justified doubt whether the choice of a digraph vs. a single vowel rune bears reliable testimony to the vocalic segment it represents (Antonsen 1998: 156; Nielsen 2012: 55–58, Goering 2023: 166, 172). 25 There is even an uncertainty, whether the reading itself should start out as faih-: “alternatively, we could view the first stave separately from the rest and identify it as i” (Zilmer & Vasshus 2023: 273). 23 Page 13 of 19 Accepted author manuscript, 25.11.2024, submitted to NOWELE 77.2 Finally, it is easy to agree with Schulte (2024: 52), when he in his discussion of Myrvoll invokes scholarly opinion that the Rök inscription (dated to the early 800s) cannot be considered URNORDISCH, i.e. that few scholars would approve of classifying its language as such. Yet, from the fact that this language is different from its ancestor it does not follow that the limit should be set at a graphemic juncture of 700/750 rather than at a phonological one. Using the end of the first syncope period is likely to result in an earlier delimitation, say 650/700. Also, the difference between the language of Rök and the classic URNORDISCH does not remove the fact that the language changed again around 850. As Schulte (2024: 52) readily admits, the SYNCOPE ERA was not over before that. At that point in time, the loss of second syllable -u after light syllables may not have been the only change, but the shortening of remaining long unstressed vowels may also belong to the ninth century (Goering 2023: 192–193). In my view, a phonological rather than graphemic argument can be mounted for at least two major structural changes around 650/700, i.e. in slight but significant anticipation of the runic transformation. First, this is likely to have been the period when fricatives lost their voicing distinction in non-initial positions (Barnes 2004; Schulte 2018a: 44), Second, a case has been made for a delay in the syncope of /u/ after a heavy syllable (Grønvik 1998: 19–21), and this led to a structurally important round of labial umlaut (Schalin 2024). In all, these considerations justify abandoning the alphabetic criterion and delimiting a subperiod from 650/700 up until ca. 850, a stage during which the umlauts were generalised variably in different dialects and new vowel loss was temporarily paused. 3.2 Observations on Goering (2023) and the syncope era The monograph by Nelson Goering (2023) is an important work for the discussion of syllable weight and foot structure in Old English and early North Germanic. It brings together arguments from poetic meter and phonological moraic theory, with Chapter 10 examining thoroughly and critically the potential testimony of runic inscriptions for early North Germanic. Whereas this broad and profound perspective, highly relevant for periodisation, has become unusual in the literature (Schalin 2018: 71– 73), Goering places himself at the forefront of bringing together the arguments from historical phonology and traditionalist philology on such issues. Goering’s analysis comes out in support of a well-known fact: the first wave(s) of syncope were significantly earlier than the last wave around 850. He disagrees with some who have viewed the length of the interregnum as a problem of plausibility: in West Germanic the equivalent pause has been immensely longer (Goering 2023: 187–188). Further, Goering (2023: 173–174; 186–187) is cautiously inclined to accept Kiparsky’s (2009: 19–26) argument for sequencing the first syncope period differently than done thus far. A first wave would have eliminated unstressed vowels only where deletion would not create overheavy initial syllables. The runic evidence from the earlier of the Blekinge stones testifies to a stage, where the constraint against over-heavy syllables was satisfied by epenthetic vowels in other positions (Goering 2023: 169–174). What neither Kiparsky nor Goering discusses, is how the quality (sonority) of the vowels targeted for loss might have affected the chronology. It has been argued that explaining the idiosyncratic outcomes of labial and front umlaut require more than two consecutive restructurings of the vowel inventory, for which a later loss of -u than -i, or indeed -a, may be a reason (Skomedal 1980: 124-126; Myrvoll 2012: 25–28; Schalin 2017: 192–193; cf. Grønvik 1987:167-189, 1998: 16–26). Further, Goering (2023: 176–177, 180–181) makes a good case for rejecting Kiparsky’s (2009: 26) minority-view hypothesis that CV̆CV̆C words would have lost their second syllable already in the first syncope period.26 However, when Goering (2023: 177–178) expresses reluctance to accept “Kiparsky Goering’s (2023: 176) reference “(so also Riad 1992: 108–118)” could give the impression that Riad’s position on the issue of early syncope would be in line with Kiparsky’s, even if Riad (like Goering) argues for late loss of 26 Page 14 of 19 Accepted author manuscript, 25.11.2024, submitted to NOWELE 77.2 and Riad’s ideas of early medial syncope” of words beginning with the structure CV̆CV̆CV:- he is less assertive and less convincing. Goering (2023: 177) clearly admits that (for hindering medial syncope) the “the minimal-word constraint would not be an issue”. On final consonant extrametricality, which Goering takes an issue with, the argument would here have been easier to follow if Kiparsky and Riad had been referenced separately and in more detail, as their analyses differ on this point.27 When Goering (2023: 178) sums up he introduces “theoretical grounds” for doubting whether ideas of early medial syncope can be sustained, but his formulation does not guide the reader to a clear understanding of what these theoretical grounds are and why syncope should not be expected here. When he states his doubt about whether “Kiparsky and Riad’s ideas of early medial syncope can be sustained, resting as they do on untenable assumptions about general final-consonant extrametricality”, the reader must connect the dots and consult the sources to infer what Goering may mean.28 Here forecasting subsection 10.3.6. would have been useful with an indication of what may have hindered syncope medially, such as perhaps protection afforded on account of the syllable belonging to the main stress foot, prior to “a shifting preference to align feet and syllables more closely” (Goering 2023: 182). Goering (2023: 177–178, 180) compiles and analyses the runic evidence relevant to medial syncope of CV̆CV̆CV:- words. To make his point he must find various ways to explain five different runic occurrences where vowel deletion seems to have taken place early. While he does this by four very different and variably strained lines of argument, he does not address wate *[wǣtē] (< *watijē) in the Strøm inscription (KJ 110), perhaps since its morphological interpretation is somewhat disputed. Goering is also open to dating medial syncope as marginally earlier in the context of the second syncope period, rather than antedating it to the first. This would remove the need to explain three syncopated examples from the Rök inscription. Notwithstanding, the remaining explanatory burden seems excessive, including for nᴀkdąn (< *nakwidan) in the Eggja inscription. 4 Reflections and proposals on periodisation and labelling of early North Germanic language Schulte’s initiative to discuss periodisation provides an excellent opportunity to seek convergence on a coherent periodisation and labelling practice. One opportunity was already lost after the publication of Nielsen (2000) and the opening discussions in Bandle et al. (2002). In order to achieve a consensus, I believe the following insights and elements must be taken into account. 1. The periodisation of early North Germanic language should support reconstructive work in historical linguistics and therefore not be in conflict with the structurally most relevant linguistic events that may delimit levels of reconstruction. /i/ and /u/ in CV̆CV̆C words. I take it to mean that Riad’s position is similar to Kiparsky’s on the statement of the last subclause “just so long as the word maintained an overall minimum of two moras”. More importantly in this context, Riad (1992: 141–143) nonetheless views the syncope to be late, which is explained by the ordering of a foot assignment rule before the mora deletion rule. 27 It is a basic tenet of Kiparsky’s analysis, while Riad (1992: 43–44; 105–106, 140) holds some word-final sonorants for moraic, notably at least those in syllables other than the main-stressed one. Another difference is that in Kiparsky’s analysis (unlike Riad’s) medial syncope after a heavy stem would have been somewhat delayed as it creates an overheavy syllable, as in *døø.mi.doo > *døøm.doo ’judged, deemed’. 28 It seems to mean that using final-consonant extrametricality to explain a delay in other instances exemplified by sunu > sun implies syncope by default medially, since the constraint would not apply and thus would not block. Page 15 of 19 Accepted author manuscript, 25.11.2024, submitted to NOWELE 77.2 2. The periodisation of runic typology must not be confused with the periodisation of language stages. When runic interpretations are used as testimony to language, and vice versa, due diligence must apply. 3. The language stages should not be labelled based on the writing system used, nor on other socio-cultural criteria. 4. Many structural changes relevant to diachronic phonology seem to cluster around the termination of the earlier syncope period(s), some time before the heyday of rune-typological transformation. It is logical to delimit the language stage from this point to the second syncope period as a language phase in its own right, with a label of its own (e.g. EARLY OLD NORDIC). It is also a period for which research with the comparative method based on OLD ICELANDIC on the one hand, and various eastern varieties on the other, holds a promise, since the eastern vowel systems cannot be derived from the OLD WEST NORDIC one (Schalin 2024). 5. The life span of the Elder Futhark 1/50–700/750 does not offer a perfect fit as a linguistic period. At the end of the timeline, the use of the Elder Futhark most likely infringes on the language stage mentioned in the previous bullet point, represented in the Eggja inscription (KJ101) still relying on transitional runes from the elder futhark. The rise of runic writing also likely infringes on the NORTHWEST GERMANIC language stage at the other end of the timeline, even if not quite as much as previously asserted by some. Still the burden of proof lies on whoever wants to argue that earliest Runic could show enough North Germanic innovations at different levels of grammar in the first (or even second) century CE. 6. The famously dynamic first syncope period is likely to have progressed in at least two or three consecutive restructurings, separated by at least as many generations of language learners. To facilitate a scholarly discussion on these stages, as for example in Grønvik (1998: 16–26), Myrvoll (2012: 25–28), Schalin (2021: 25–27) and Goering (2023: 184–186), there is a need to delimit this period as a main period even if it is short, unstable and hard to date precisely. Consequently, temporal qualifiers such as “Early/Elder” or “Late/Younger” can rather be dispensed with, since these terms should be reserved for even more fine-tuned analytically necessary distinctions (see note 15). This point cannot be stressed enough. The use of TRANSFORMATIONAL SCANDINAVIAN would avoid confusion with the established terminology of TRANSITIONAL inscriptions. 7. The stable language stage with “mildly Scandinavian” (Haugen 1976: 127; Nielsen 2000: 287– 293) features that preceded TRANSFORMATIONAL SCANDINAVIAN and followed NORTHWEST GERMANIC needs a label which is not a misnomer. Since this stage ended earlier than what some refer to as ANCIENT NORDIC, it should not assume that name.29 While it is older than the “proto-language” achievable by the comparative method, the label PALAEO-SCANDINAVIAN seems appropriate, in analogy with the term PALAEO-GERMANIC occasionally used for the language stage preceding reconstructed PROTO-GERMANIC (Kallio 2012: 228 nt 4). It is possible that a stage of fifth century late PALAEO-SCANDINAVIAN phonology can be internally reconstructed pending future umlaut research. 8. The highly ambiguous use of the term OLD NORSE creates absurd and confusing terminological conflicts (see discussion referencing note 10). We need a convention to consistently use OLD EAST and/or WEST NORDIC respectively for the period that falls between the later syncope period and the year 1050/1100, which quite commonly has been in use to mark the emergence of the nationally designated languages. The time after this naturally falls outside the scope of international standardisation. Another disadvantage is the minimal distinction it creates to the temporal qualifier most used for the medieval languages, which is OLD. In common language these two are nearly synonymous, creating confusion. 29 Page 16 of 19 Accepted author manuscript, 25.11.2024, submitted to NOWELE 77.2 A possible periodisation based on these is illustrated in Figure 2. Note that the illustration aspires, albeit in a very tentative way, to reflect the discussion on dialect diversification during the Old Nordic period, which references note 8 in subsection 2.1. Figure 2. An illustration of a possible periodisation and labelling of early North Germanic, for further debate Acknowledgements I thank Professor Michael Schulte (UiA University of Agder) and Dr. Nelson Goering (Postdoctoral Fellow, University of Oslo) for their relevant and highly useful comments. I am also indebted to two anonymous reviewers, who contributed substantially to several points in the argumentation. Responsibility for the text, especially for potential faults or deficiencies, remains with the author. References Antonsen, E. H. 1975. A concise grammar of the Older Runic inscriptions: Tübingen: Niemeyer. Antonsen, E. H. 1994. The earliest attested Germanic language, revisited. NOWELE 23. 41–68. https://doi.org/10.1075/nowele.23.02ant Antonsen, E. H. 1998. On runological and linguistic evidence for dating runic inscriptions. In K. Düwel & S. Nowak (eds.). Runeninschriften als Quellen interdisziplinärer Forschung: Abhandlungen des Vierten Internationalen Symposiums über Runen und Runeninschriften in Göttingen vom 4.-9. August 1995 = Proceedings of the Fourth International Symposium on Runes and Runic Inscriptions in Göttingen, 4-9 August 1995, 150–159. Berlin: De Gruyter. Bammesberger, A. 2008. Runic names in -warijaz. NOWELE 53. 13–18. https://doi.org/10.1075/nowele.53.02bam Page 17 of 19 Accepted author manuscript, 25.11.2024, submitted to NOWELE 77.2 Bandle, O. 2002. Diachrony and synchrony in Nordic language history. In O. Bandle et al. (eds.) 2002. The Nordic languages. An International Handbook of the History of the North Germanic Languages. Volume 1, 25–30. Berlin: De Gruyter. Bandle, O. 2011. Die Gliederung des Nordgermanischen. Reprint der Erstauflage mit einer Einführung von Kurt Braunmüller. (Beiträge zur Nordischen Philologie, 47). Tübingen und Basel: A. Francke Verlag. Bandle, O. et al. (eds.) 2002. The Nordic languages. An International Handbook of the History of the North Germanic Languages. Volume 1. Berlin: De Gruyter. Barnes, M. P. 1998. The transitional inscriptions. In K. Düwel & S. Nowak (eds.). Runeninschriften als Quellen interdisziplinärer Forschung: Abhandlungen des Vierten Internationalen Symposiums über Runen und Runeninschriften in Göttingen vom 4.-9. August 1995 = Proceedings of the Fourth International Symposium on Runes and Runic Inscriptions in Göttingen, 4-9 August 1995, 448–461. Berlin: De Gruyter. Barnes, M. P. 2004 Spirant denotation by younger fuþark b. In A. van Nahl, L.Elmevik & S.Brink (eds.), Namenwelten. Orts- und Personennamen in historischer Sicht, 605–614. Berlin: De Gruyter. https://doi.org/10.1515/9783110911473.605 Birkmann, T. 1995. Von Ågedal bis Malt. Die skandinavischen Runeninschriften vom Ende des 5. bis Ende des 9. Jahrhunderts. Berlin: De Gruyter. Birkmann, T. 2002. From Ancient Nordic to Old Nordic: definition and delimitation of the period. In O. Bandle et al. (eds.) 2002. The Nordic languages. An International Handbook of the History of the North Germanic Languages. Volume 1, 691–698. Berlin: De Gruyter. Boutkan, D. 1995. The Germanic ‘Auslautgesetze’. Amsterdam: Rodopi. Bull, T. & M. Schulte et al. 2018. Forord til verket. In H. Sandøy, & A. Nesse (eds.), Norsk språkhistorie: Tidslinjer, 7–11. Oslo: Novus. ‘Category: Old Swedish language’. 2024. Wikipedia. Available at https://en.wiktionary.org/wiki/Category:Old_Swedish_language Fairfax, E. 2015. The runic inscription from the Nydam axehandle. NOWELE 68. 153–181. https://doi.org/10.1075/nowele.68.2.02fai Fulk, R. D. 2018. A Comparative grammar of the early Germanic languages. Amsterdam: Benjamins. https://doi.org/10.1075/sigl.3 Goering, N. 2023. Prosody in Medieval English and Norse. Oxford: Oxford University Press. https://doi.org/10.5871/bacad/9780197267462.001.0001 Grønvik, O. 1987. Fra Ågedal til Setre. Sentrale runeinnskrifter fra det 6. Århundre. Oslo: Universitetsforlaget. Grønvik, O. 1998. Untersuchungen zur älteren nordischen und germanischen Sprachgeschichte. Frankfurt am Main etc: Peter Lang. Haugen, E. I. 1976. The Scandinavian languages: An introduction to their history. Cambridge, MA: Harvard University Press. Haugen, E. I. 1982. Scandinavian language structures: A comparative historical survey. Tübingen: Niemeyer. Haugen, O. E. 2012. Frå urnordisk til norrønt språk. Arbeidsmanuskript 30. desember 2012. Retrieved from http://folk.uib.no/hnooh/handbok/ekstra/urnordisk.html Page 18 of 19 Accepted author manuscript, 25.11.2024, submitted to NOWELE 77.2 Jørgensen, B. 2016. Sproghistoriske perioder. In E. Hjorth et al. (eds), Dansk Sproghistorie 1. Dansk tager form, 79–83. Copenhagen & Aarhus: Det Danske Sprog- og Litteraturselskab & Aarhus Universitetsforlag. Kallio, P. 2012. The Prehistoric Germanic Loanword Strata in Finnic. In R. Grünthal & P. Kallio (eds.), A linguistic map of prehistoric Northern Europe. Helsinki: Société Finno-Ougrienne. Karker, A. 1996. Politikens sproghistorie. Udviklingslinjer før nudansk - Lydhistorie, bøjningshistorie, syntakshistorie og ordhistorie. Århus: Politikens Forlag. Kiparsky, P. 2009. Syncope, umlaut, and prosodic structure in early Germanic. Manuscript, Stanford University. Retrieved from https://web.stanford.edu/~kiparsky/Papers/weakpreterite.2006b.pdf Kuhn, H. 1955. Zur Gliederung der germansichen Sprachen. Zeitschrift für deutsches Altertum unddeutsche Literatur 86. 1–47. Luobbal Sámmol Sámmol Á. (Aikio, A.). 2020. Loanwords from Unattested Nordic Source Forms in Saami. Finnisch-Ugrische Forschungen 65. 5–24. https://doi.org/10.33339/fuf.87404 Myrvoll, K.-J. 2012. Framvoksteren av sterkt hokyn frå urnordisk til norrønt. In J. O. Askedal, T. Schmidt, R. Theil, & H. Bjorvand (eds.), Germansk filologi og norske ord: festskrift til Harald Bjorvand på 70-årsdagen den 30. juli 2012, 16–32. Oslo: Novus Forlag. Myrvoll, K.-J. 2022. [Rezension von] Schulte 2018a. In Futhark 12 (2021, publ. 2022), 183–197. https://doi.org/10.33063/diva-491890 Nedoma. R. 2010. Kleine Grammatik des Altisländischen. 3. Aufl. Heidelberg: Winter. Nesse, A. 2018. Periodeinndeling, kildetilfang og språkhistoriske innfallsvinkler. In H. Sandøy, & A. Nesse (eds.), Norsk språkhistorie: Tidslinjer, 33–49. Oslo: Novus. NSH IV = Sandøy, H. & Nesse, A. (eds.). 2018. Norsk språkhistorie: Tidslinjer. Oslo: Novus. Nielsen, H. F. 2000. The early runic language of Scandinavia: studies in Germanic dialect geography. Heidelberg: Winter. Nielsen, H. F. 2002. Delimitation of Ancient Nordic from Common Germanic and Old Nordic. In O. Bandle, et al. (eds.) 2002. The Nordic languages. An International Handbook of the History of the North Germanic Languages. Volume 1, 615–618. Berlin: De Gruyter. Nielsen, H. F. 2012. Gothic Runic Inscriptions in Scandinavia? Futhark 2. 51–61. Ralph, B. 1980. Urnordisk slutljudsskärpning. Arkiv för Nordisk Filologi 95.1–24. Riad, T. 1992. Structures in Germanic prosody. A diachronic study with special reference to the Nordic languages. Ph.D. dissertation. Stockholm: Dept. of Scandinavian languages, Stockholm University. https://urn.kb.se/resolve?urn=urn%3Anbn%3Ase%3Asu%3Adiva-14475 Ringe, D. 2006. From Proto-Indo-European to Proto-Germanic. A Linguistic History of English, Volume I. Oxford: Oxford University Press. Rischel, J. 2008. A Unified Theory of Nordic i-Umlaut, Syncope and Stød. NOWELE 54/55. 191–235. https://doi.org/10.1075/nowele.54-55.06ris Schalin, J. 2017. Scandinavian umlaut and contrastive feature hierarchies. NOWELE 70. 171–254. https://doi.org/10.1075/nowele.70.2.03sch Schalin, J. 2018. Preliterary Scandinavian sound change viewed from the east: Umlaut remodelled and language contact revisited. Summarising chapter of compilation thesis, Helsinki: Helsinki University. http://urn.fi/URN:ISBN:978-951-51-4387-7 Page 19 of 19 Accepted author manuscript, 25.11.2024, submitted to NOWELE 77.2 Schalin, J. 2021. Nordic umlaut, contrastive features and stratal phonology. In Nordlyd 45.1, selected papers from the fifth Fonologi i Norden Meeting, edited by Islam Youssef and Miguel VázquezLarruscaín, 7–37. https://doi.org/10.7557/12.6249 Schalin, J. 2024. North Germanic labiovelar breaking – reconstructing the phonological conditions. Presentation given on 24 Febuary 2024 at the conference Fonologi i Norden (FiNo2024), Stockholm. https://doi.org/10.13140/RG.2.2.21465.49762/1 Schmidt, Ch. G., Nedoma, R. & Düwel, K. 2012. Die Runeninschrift auf dem Kamm von Frienstedt, Stadt Erfurt. Die Sprache 49(2) (2010/2011). 123–186. Schulte, M. 1998. Grundfragen der Umlautphonemisierung. Eine strukturelle Analyse des nordgermanischen i/j-Umlauts unter Berücksichtigung der älteren Runeninschriften. Berlin: De Gruyter. Schulte, M. 2018a. Urnordisch. Eine Einführung. Wiener Studien zur Skandinavistik 26. Wien: Praesens-Verlag. Schulte, M. 2018b. Vikingtiden (700–1050). In H. Sandøy, & A. Nesse (eds.), Norsk språkhistorie: Tidslinjer, 119–196. Oslo: Novus. Schulte, M. 2024. Stationen der frühen nordgermanischen Sprachgeschichte: Zu einer neuen Periodisierung des Urnordischen. NOWELE 77(1). 49–80. https://doi.org/10.1075/nowele.00084.sch Schulte, M. & H. Williams. 2018. Den eldste tiden (–700). In H. Sandøy, & A. Nesse (eds.), Norsk språkhistorie: Tidslinjer, 51–117. Oslo: Novus. Skomedal, T. 1980. Syncope, omlyd og brytning i nordisk. In E. Hovdhaugen (ed.), The Nordic languages and modern linguistics: Proceedings of the Fourth International Conference of Nordic and General Linguistics in Oslo 1980, 120–139. Oslo: Universitetsforlaget. Skre, D. 2022. Scandinavian kingship AD 500-1000. In Ludowici, B. and Pöppelmann, H (eds.) New Narratives for the First Millennium AD? : Alte und neue Perspektiven der archäologischen Forschung zum 1. Jahrtausend n. Chr. (Neue Studien zur Sachsenforschung 11), 117–131. https://doi.org/10.24355/dbbs.084-202301131148-0 ‘Spoken L1 Language: Old Norse’. 2024. Glottalog 5.0. https://glottolog.org/resource/languoid/id/oldn1244 Syrett, M. 1994. The Unaccented Vowels of Proto-Norse. Odense: Odense University Press. https://doi.org/10.1075/nss.11 Venås, K. 2002. Previous attempts at establishing periods in Nordic language history. In O. Bandle et al. (eds.) 2002. The Nordic languages. An International Handbook of the History of the North Germanic Languages. Volume 1, 31–37. Berlin: De Gruyter. Voyles, J. B. 1992. Early Germanic grammar: pre-, proto-, and post-Germanic languages. San Diego, CA: Academic Press. Widmark, G. 2004 [2001]. Det språk som blev vårt: ursprung och utveckling i svenskan: urtid, runtid, riddartid (reprint with minor corrections of 1. ed. from 2001). Uppsala: Kungl. Gustav Adolfs akademien för svensk folkkultur. Zilmer, K. & K. Vasshus. 2023. Runic Fragments from the Svingerud Grave Field in Norway: Earliest Datable Evidence of Runic Writing on Stone. NOWELE 76(2). 233–303. https://doi.org/10.1075/nowele.00080.zil