Exploring Variation and Change in the Distant Past

Lay Summary

Much early work in how languages present variation related to language change started with studies of European languages in the Early Modern period into the 19th century. This paper explores a greater time depth and a part of the world where we have less of this kind of work to date. The study documents change in sibilant consonants in Old Babylonian letters, a type of sound change, and a data source widely used in historical sociolinguistics. While current work increasingly goes farther back into history and covers settings around the world, this contribution shows just how much we can learn about linguistic variation and change and how deep into the past we can go.

Introduction

The exploration of linguistic features for languages of the past is a challenging task. We can go back a long way in time to investigate some general dynamics of linguistics features prior to their direct attestation. The modelling of language traits through probabilistic comparative analyses and reconstructions offers a unique insight into such pre-documented language dynamics in the distant past (e.g., Dunn et al., 2011; Carling; Cathcart, 2021). Nevertheless, to analyse the particulars of the process of language change, the co-existence of competing linguistic forms in language variation, and their socio-historical embeddings, we are constrained to analysing attested forms of language usage. For the vast majority of our past ―before the earliest audio-recordings in the second half of the 19^th century― such attested language comes in the form of written documentation. Spoken language is of course not identical to written language, but the latter remains a linguistic domain which has the ability of capturing fossilized language communication from afar spaciotemporal settings.[1] In as much as it represents a linguistic phenomenon, it can be argued that written language (and the variation and change it preserves) deserves linguistic research, as does spoken language (Romaine, 2009, p. 14–21).

The field of historical sociolinguistic research deals with the aforementioned questions ―as well as others, such as the conscious or unconscious acts of social identity represented in written records (Auer et al., 2015, p. 9)― by analysing diverse corpora of historical languages. The field initially focused mostly on well attested post-medieval European languages (e.g., Nevalainen; Raumolin-Brunberg, 2003; Rutten; van der Wal, 2014) but, as the present special issue demonstrates, its scope has rapidly become much wider. However, not all written records are equally suitable for this endeavour. In diachronic terms, it is reasonable to expect major limitations in sociolinguistic approaches to ancient languages. This leads to the question: how far back in time can we go to study variation and change in languages?

Some notable studies of classical languages have swiftly followed the (historical) sociolinguistic route, assessing how social-historical cues relate to the preserved record of languages such as Latin or ancient Greek (e.g., Adams, 2013; Clackson, 2015; Mullen; James, 2012).[1] By contrast, ancient languages are often seen as too far removed from sociolinguistic practices and too fragmentary, formulaic or uncertain to allow for informed assessment on variation. An illustrative example of the justified scepticism about sociolinguistic research on ancient languages at the turn of the century concludes that “a more than anecdotal sociolinguistic interpretation of linguistic data of a remote past is beset with near-insurmountable difficulties.” (Winter, 1999, p. 79).[2] Happily, the available data and the capabilities of present-day computer technology are notably enhanced compared to those of past research efforts.

For ancient languages, it is also worth considering what a researcher might consider achievable. Whilst the aims of a sociolinguistic project working on, say, present-day Javanese, would inevitably differ from those of researchers looking at data obtained from XVIII century Dutch letters, for languages further down the timeline the level of expectations must correspondingly have a lower resolution. Determining whether insights into the co-occurrence of linguistic and extralinguistic factors is relevant for early attested languages depends on this scale. On the other hand, there are still requirements that need to be fulfilled in order to achieve reliable results. As Winter rightly pointed out, sociolinguistic conclusions about languages of the past depend on “the availability of extralinguistic information and on the nature of the texts one can use” (Winter, 1999, p. 79).

A complementary expectation regarding languages of the distant past is that they would have behaved in a fundamentally similar way to languages of the present. This common assumption is one of the various interpretations of the otherwise loose notion referred to as the uniformitarian principle (Walkden, 2019). The principle plays an important role in various branches of linguistics, notably in historical linguistics and in typology (Song, 2018, p. 22–26). From a sociolinguistic perspective, the patterns in which languages vary in present-day languages are expected to be similar to those that operated in the past (Romaine, 1988, p. 1454), though particular attention must be taken not to engage in anachronistic assumptions regarding the social structures of past communities (Bergs, 2012).[1] Labov’s take on uniformitarianism gives some defined temporal parameters: “the forces operating to produce linguistic change today are of the same kind and order of magnitude as those which operated five or ten thousand years ago” (Labov, 1972, p. 275) [emphasis added]. The present article presents data that is not quite as remote as Labov’s reference but is not far from it: it dates back almost four thousand years. It is indeed closer to the time horizon mentioned by Labov than most other attested languages, and its antiquity doubles the temporal span that separates XVIII century Dutch letters in Rutten & van der Wal (2014) from Latin texts in Adams (2013). Despite its age, this linguistic data are quantitative and qualitatively valuable and can be informative for research on language variation and change.

The following section (2) introduces the characteristics of Old Babylonian and its textual record. It also presents some studies on variation in languages that are even older than the Old Babylonian record. Section (3) focuses on the study of a particular variable in a corpus of Old Babylonian letters, presenting data from Hernáiz (2020) and confronting it with new data extracted from an additional group of 219 letters. The last section (4) offers some conclusions, bringing the results of the case study to the general discussion points introduced here.

1. Language variation (and change) in the distant past: Old Babylonian

The reference to a historical juncture as distant is, of course, inherently relative. Its significance is contingent on specific temporal parameters or, more broadly, on our perspective within a given context. At the risk of giving the reader the wrong impression that the article is about some time in the early Holocene (or before), the deliberate choice of the term distant in the title underscores the reference being made to the earliest period with recorded evidence of language variation and change. Additionally, the term serves the purpose of highlighting the antiquity of the temporal frame discussed here, which is not sufficiently stressed by terms such as ancient. Western conventional segmentations of the past into ancient history, medieval history, modern history, etc., do not correspond to uniform temporal intervals and may inadvertently obscure the true temporal extent of each period. While Classical Latin and early Akkadian are both considered ancient languages (i.e., spoken in antiquity), the distance between their respective written evidence is about two thousand years, in turn, the same distance that separates Classical Latin from present-day languages. On the other hand, the absolute chronology of the Old Babylonian record shows that it is in fact contemporary to events that are typically perceived as distant in time. For example, while Old Babylonian texts were written in western Asia, contemporaneously, people in Southern Britain were engaged in the final stages of constructing the Stonehenge complex. In most other parts of the world, this same period effectively qualifies as prehistory, implying that the languages spoken there and then can only be inferred through reconstructions based on comparative analyses of related documented languages. Old Babylonian is temporally distant enough to be dated to an intermediate period between some of these proto-languages, such as Proto-Indo-Iranian and Proto-Indo-Aryan.

Whilst Old Babylonian (Semitic: Iraq: ca. 4000-3500 YBP) stands as one of the earliest documented languages, it is by no means the oldest language record for which some kind of variation can be established.[1] Nevertheless, the focus of the present study on Old Babylonian is simply driven by the sheer volume of textual evidence available for this language, offering an exponentially greater amount of texts compared to older attested languages. The wealth of data allows for a more comprehensive documentation of (written) language use and its variation and change with a surprising high-level resolution, considering its antiquity. Old Babylonian was profusely used in written documents of different genres (administrative, letters, literary, scientific, monumental etc.) over a region comprising most of today’s Central and Southern Iraq and adjacent regions in Syria. Current estimations suggest that around 50,000 Old Babylonian texts are known to scholars (Streck, 2022, p. 2),[2] making OB the best attested variety of the overarching linguistic clade that we call Akkadian language. It should be mentioned that Akkadian ―unlike other ancient languages such as Ancient Greek, Latin or, to a lesser extent, Hittite― remains a language relatively unexplored by general linguists. And this is despite its large textual record and the fact that it is one of the few languages for which documentation spreads fairly consistently over two millennia (with its potential for historical linguistic research). According to the website of the Department of Near Eastern Languages & Civilizations at Yale University, the available corpus of Akkadian (more than half a million documents!) is “larger than that of classical Latin and surpassed in size by only one ancient language, Greek”.[3] The volume of documentation and our growing knowledge about its historical embeddings, therefore, make Akkadian in general and Old Babylonian in particular, reasonable candidates to meet the above-mentioned conditions for conducting historical sociolinguistic research in ancient languages.

On the other hand, historical sociolinguistic research on Old Babylonian is unsurprisingly fraught with methodological and documentary challenges. To the limitations of any historical sociolinguistic study (such as the need to work exclusively on written language, the accidental nature of the record, the lack of representativeness for certain groups, etc.) we need to add the fragmentary state of part of the documentation, the uncertainties about the specific temporal and geographical coordinates from which the documents stem, the elusive information concerning the social structure of ancient Mesopotamia, the question of scribal education and the potential use of writing by non-professional scribes. Nonetheless, the case study presented in the next section shows the results of analyses on a corpus of ca. 1800 Old Babylonian letters[2] for which certain language-external information exists regarding the individuals, the time, or the region associated to the production of the texts.[3] The representability of the data in the corpus is constrained by the unbalanced proportion of letters that relate to specific areas, periods or individuals, an issue that is naturally related to the intricacies of the archaeological findings and the publication of texts’ editions. But although the record is accidental and uneven, the use of various types of data from interdisciplinary sources can be applied to Old Babylonian ―with the appropriate caveats― just as it is indeed applied elsewhere in historical sociolinguistic research (Lauersdorf, 2021). A truly interdisciplinary approach that incorporates any information (linguistic or else) is, of course, more informative. In spite of the limitations and potential inadequacies in text classification, a comparison of a large volume of textual material unveils cohesive patterns that support the overall validity of the classification of texts in certain spatiotemporal categories. Although most texts analysed in section 3 contain no direct reference to a specific date, some of the writers (or rather the senders of the letters) can be identified by parallel archives. In other instances, allusions to historical figures or archaeological data provide crucial clues for approximating a general chronological timeframe. A notable convergence of orthographic and linguistic variants tends to occur in specific segments of the corpus, implying that the criteria utilized for the regional and temporal taxonomy of the texts has a nontrivial degree of relevance for a preliminary categorisation (Hernáiz, 2020).

Akkadian varieties neighbouring Old Babylonian are also well-known (especially Old Assyrian), but the present study focuses on exploring the internal variation within documents from the core Old Babylonian area of Mesopotamia (i.e., excluding other variation documented in ‘peripheral’ areas such as the Middle and Upper Euphrates basin or the Khabur area of modern Syria).[1] Hence, the restricted object of study here is the central Babylonian variety (commonly regarded as its ‘classical’ form), which forms the basis for most grammars of Akkadian, and has been typically regarded as a sort of standardised chancery variety from the time of King Hammurabi of Babylon (Soden, 1995, p. 3). Against this view, Westenholz (2006) describes the language of this period as “a language in lively development […] without any fixed written norm” (Westenholz, 2006, p. 254). This core Mesopotamian area is a fertile plain of about 310 km on a straight line from the southern city of Ur to the northern city of Eshnunna,[2] delimited by desert to the west and mountain ranges to the east. It was communicated by the course of the rivers Euphrates and Tigris plus a well-developed net of canals. The region, however, was not socio-politically uniform. Throughout the OB period (particularly in the first half) several state entities controlled their own territory and resources, often engaging in hostilities or alliances with their neighbours. The texts of the corpus are conventionally divided into three categories, according to three broad regions commonly distinguished in studies on Old Babylonian documents: a southern region (the area that had been the homeland to Sumerian speakers before this period), a northern region (with Babylon as its centre) and the valley of the Diyala river, near modern-day Bagdad.

Figure 1. Figure 1. Map of central modern Iraq showing the location of Old Babylonian centres: Shaduppum (in grey, Diyala region), Babylon (in black, northern region) and Larsa (in white, southern region). v Map based on Urutseg, licensed under CC0, via Wikimedia Commons, modified by Rodrigo Hernáiz.

2. Case study: Old Babylonian /s/ in space and time

Despite the fact that Akkadian ceased to be spoken about 2000 years ago, two centuries of philological scholarship and the relatively straightforward comparison of Akkadian texts with other better-known Semitic languages have allowed a good understanding of its typology, lexicon and general phonological traits. The present study seeks to investigate the written realization of a phoneme, conventionally rendered /s/, in the corpus of Old Babylonian letters. But is it possible to detect any signs of sound variation in a language whose record predates the earliest alphabetic scripts?

The cuneiform writing system used in the production of Old Babylonian texts consisted of different types of signs. The most important ones are logograms (word-graphemes with no direct relation to sounds) and syllabograms, which represent combinations of consonants and vowels.[1] Fortunately, the OB epistolary genre exhibits a greater use of the latter, compared to other text genres. This study focuses exclusively on the assessment of consonant-vowel syllabograms with the onset consonant /s/. This phonemic unit is in fact a conventional way to refer to a sound which corresponds to voiceless sibilant phonemes in Semitic cognates. However, the realisation of this notional phonemic unit was probably not uniform throughout diachronic (or diatopic) varieties of Akkadian.

In the earliest period (Old Akkadian), /s/-syllables were spelled with a set of graphemes conventionally labelled ZA, ZI and ZU (henceforth Z-signs). These signs were also used in the spelling of voiced and ‘emphatic’ sibilants.[1] Based on language-internal cues and Semitic comparative analyses, it is widely accepted that Old Akkadian /s/ was in fact an affricate [t͡s] (Faber, 1985; Sommerfeld, 1995; Streck, 2006). By the Middle Babylonian period ―about five hundred years later― voiced and ‘emphatic’ sibilants continued to be represented by Z-signs. However, /s/ was then consistently rendered by a different set of syllabograms, namely SA, SI, SU (S-signs).[2] The explanation invoked for the orthographic change is that a sound change (deaffrication) occurred at some point in the history of Akkadian, as well as in other Semitic languages (Sommerfeld, 1995). According to this, said change —a completely normal sound change, widely attested typologically in many languages — is responsible for the orthographic shift from Old Akkadian (Z-signs) to Middle Babylonian (S-signs).

But what is the situation in the intermediate period between Old Akkadian and Middle Babylonian? The answer is that Old Babylonian exhibits both sets of Z-signs and S-signs to render syllables with onset /s/. Thus, the Old Babylonian corpus contains variation and comprises a relatively wealthy number of texts, two necessary conditions for variationist research. Additionally, a growing amount of information continues to be published regarding social, economic and political issues of the time (e.g., Charpin; Edzard; Stol, 2004), or about families and individuals mentioned in Old Babylonian archives (e.g., Jacquet, 2013). A research question arises as to what a historical sociolinguistic investigation of the corpus of OB letters can tell us about this type of variation in spelling, within the context of a plausible sound change.

In the 1940’s, German scholar Albrecht Goetze published an observation on the spelling of sibilants in a group of Old Babylonian mathematical texts (Goetze, 1945). In this article and in a subsequent piece (Goetze, 1958), Goetze established the influential idea of a dialectal distinction in Old Babylonian, whereby northern areas of the core Mesopotamian milieu were characterised by the use of S-signs to represent /s/, while southern areas exhibit Z-signs in the same environments. This geographic explanation for the variation has been tacitly challenged by the more recent observation that the choice of signs may be dependent on the phonetic context in which /s/ is embedded. This new explanation was articulated in Sommerfeld (1995) and has proven to be consistent in a number of Old Babylonian texts (Streck, 2006). The general pattern observed can be summarised as follows:

· Z-signs are used in word initial and in double /s/ (either geminated or resulting from the contact between a stem-final sibilant or dental and a bound pronoun with onset sibilant). For example:

sakāpum (‘to push’) would appear written with the sign ZA (as word-initial): ZA-KA-PU-UM

issakip (‘he was pushed’) would also occur with ZA (as double /s/): IZ-ZA-KI-IP

· Non-initial short /s/ would be rendered by S-signs:

isakkip (‘he pushes’): I-SA-AK-KI-IP

This pattern strongly suggests that the eventual change in writing practices (from Old Akkadian Z-signs to later Babylonian S-signs) was ultimately motivated by a sound change in spoken usage. The idea behind is that a possibly deaffricate articulation would be first perceived mostly in intervocalic environments, whereas its (possibly affricate) allophone would still be articulated or perceived in word-initial or double instances. This is not to say that written features correspond unequivocally and in complete synchrony to spoken sounds, consistently mirroring their variation. However, the eventual shift in the writing of the phoneme conventionally denoted as /s/ seems to be motivated ―asynchronously as this should be― by a sound change which at some point affects certain phonetic environments more prominently than others.

In any case, how does this phonologically motivated pattern align with the diatopic variation suggested by Goetze? Are we to conclude that OB mathematical texts adhere to different scribal traditions in rendering /s/ compared to other genres? The analysis of the large corpus of OB letters sheds light on these questions. The bulk of the letters included in the Annotated Corpus of Correspondence in Old Babylonian (ACCOB) shows that the general distribution of consonant-vowel syllabograms representing /s/ by means of Z- and S-signs is predominantly consistent with the ‘allophonic pattern’. However, there are also diatopic and diachronic signals that have an effect on the distribution of the data. To consider the differences in the variable (S-sign, Z-sign) from a geographical point of view, Hernáiz (2020) collects all the instances of consonant-vowel syllabograms with onset /s/ in the ACCOB corpus, controlling for the categorization of the three main areas discussed previously: northern-related letters, southern related letters and letters related to the Diyala valley. Each region presented different ratios of use of Z-signs of S-signs. The graph in Figure 2 shows the occurrence of both variants in word-initial and double /s/, i.e., the contexts where one would expect to find Z-spellings, according to the phonologically motivated pattern mentioned before.

Figure 2. Figure 2. Instances of Z- and S-signs for contexts where Z-signs would be expected by their phonological context (Hernáiz 2020).S

The data indicate a statistically significant distinction among the three areas, despite the uneven number of texts/occurrences within each of the subgroups (Pearson's Chi-squared test data for independence: X-squared = 2.623, df = 2, p-value = 5.553e-10).[1] Most striking is the distinct preference for S-signs in the Diyala area, preceding the spelling practice of /s/ in later Middle Babylonian. This particular characteristic of the Diyala texts has already been observed by Sommerfeld (2006), who speaks of a deaffrication process being discernible already in Old Babylonian texts from this region. A closer analysis of the data reinforces the peculiarity of the Diyala texts: 6 of the 16 ‘expected’ instances of Z-spellings in the Diyala group belong to a subgroup containing the very earliest OB texts in the corpus (Whiting, 1987). Unlike the rest of the Diyala group, this small early subgroup is consistent with the ‘allophonic pattern’ and shows no signs of deaffrication.

In sum, while the northern- and southern-related letters present a fairly consistent distribution of Z-spellings in word initial and double /s/, the Diyala letters (except for the very earliest ones) seem to follow an innovative writing practice that might be related to a distinct spoken regional feature. It is important to bear in mind that the geographical distinctions made in the corpus (based on distinctive socio-political realities) do not represent areas in isolation. The landscape between core Mesopotamian areas was traversable and there was constant communication between the different regions. In fact, the main centres in the Diyala are very close indeed to one of the cities that provides more documentation for the northern region (Sippar). They are roughly located on opposite sides of the outskirts of modern Bagdad. Indeed, the difference in the variable (s, z) (and other features) is attested in centres that were no more than 40 kilometres apart in some cases.

Whilst Figure 2 (above) showed the instances of word-initial and double /s/, where Z-spelling are predicted by the allophonic explanation, Figure 3 (below) presents forms with non-initial short /s/ in the corpus. This is the environment where /s/ is expected to be represented by S-signs in Old Babylonian.

Figure 3. Figure 3. Instances of Z- and S-signs for contexts where S-signs would be expected by their phonological context (Hernáiz 2020).

The graph shows again that, quantitatively speaking, the corpus contains far more S-spellings than the alternative Z-signs, as expected. However, the southern-related letters show a distribution that is significantly different from the other two regions (Pearson's Chi-squared test data for independence: X-squared = 117.8, df = 2, p-value < 2.2e-16). The initial impression is that the southern group retains more Z-signs in this phonetic environment, placing it closer to previous Old Akkadian spelling practices.

It should be noted that the internal composition of each of the groups is not even: the chronology of southern-related texts is on average older than that of the bulk of norther-related texts, because it does not include ‘late Old Babylonian’ documents.[1] To expand the data from Hernáiz (2020) and minimise the potential effect of chronology on the distribution of variants, the same variable can be now analysed in a group of 219 southern OB letters published in George (2018).[2] These letters originate from the southern Old Babylonian city of Larsa or the surrounding areas under Larsa’s control. They are dated to two different periods: 89 of them can be considered ‘early’ Old Babylonian letters (predating the reign of Rim-Sin), but 130 belong to the ‘middle’ Old Babylonian period, primarily associated to the time of Rim-Sin (1822-1763 BCE) (George, 2018, p. x-xii). The data indicate that this new corpus supports the distinctively ‘conservative’ usage of Z- and S-signs to render /s/ in southern letters, compared to other core Mesopotamian areas. Importantly, this holds true not only for early Old Babylonian letters, but also for middle Old Babylonian letters, i.e., the best documented period in ACCOB, accounting for most of Diyala and northern-related texts in the corpus (Hernáiz, 2020).

To wit, Figure 4 shows the overwhelming occurrence of Z-signs in word-initial and double /s/ (where Z-spellings are indeed expected), for both early and middle OB southern texts in George (2018):

Figure 4. Figure 4. Instances of Z- and S-signs for contexts where Z-signs would be expected, in early and middle southern OB texts in George (2018).

Crucially, Figure 5 (below) shows that Z-signs are also prominent in non-initial short /s/, where S-spellings would be expected according to the ‘allophonic pattern’. Significantly, this occurs in both early and middle OB texts:

Figure 5. Figure 5. Instances of Z- and S-signs for contexts where S-signs would be expected, in early and middle southern OB texts in George (2018).

These new data reinforce the observations in Hernáiz (2020), suggesting that the southern area exhibited a more ‘conservative’ approach in rendering /s/, as it retains, to a greater extent, the writing practice observed in earlier Old Akkadian records. This trend persists not only for earlier OB periods, but also throughout the more extensively documented middle phase of the Old Babylonian period.

Even though the variable analysed (variable (s, z)) probably has an ultimate phonological motivation ―perhaps related to a general deaffrication process― the ‘conservative’/ ‘innovative’ labelling used here inevitably refers to written language. Since there are no records of spoken Akkadian, we can only speculate about the relationship between the variation in written records and variation in spoken Old Babylonian lects. The distinct distribution of /s/-variants in the Diyala, northern and southern textual records respectively might be purely attributed to varying scribal traditions in different centres. Some of these centres might have been more conservative, but only at the orthographical level. While this remains a plausible explanation, it should be noted that other orthographic/linguistic variables do not match the conservatism or innovation trends associated with variable (s, z) in these areas. For example, although southern-related letters are more ‘conservative’ representing /s/, they are more ‘innovative’ in rendering nasalization of double stop consonants, a feature that will become prevalent in Middle Babylonian (Hernáiz, 2020, p. 267–278). They are also more ‘innovative’ than roughly contemporary texts from the Diyala and northern areas in the contraction of the pronominal form šu’āti (SG.ACC/GEN) into šâti (Hernáiz in preparation). In sum, while diatopic differences in texts might be merely the product of more or less entrenched scribal orthographic traditions, a rigorous approach should not necessarily dismiss the possibility that said differences might perhaps reflect in fact some regional variation going beyond the written domain of language.

3. Discussion: variation and change in Old Babylonian

Despite the obvious limitations ultimately arising from the great distance between modern observers and the individuals who wrote Akkadian letters, the examination of the well-documented Old Babylonian material reveals significant types of variation permeating throughout the texts. Remarkably, this is evidenced to a considerable level of detail, given the time depth. Although more general accounts of dialectal or genre distinctions can be found in the record of languages that predate Old Babylonian by a few centuries, the higher resolution data in Old Babylonian corpora are particularly apt for historical sociolinguistic enquire.

Of course, these investigations do not meet far-fetched expectations about detailed questions on defined sociolects or the potential reasons why an individual would choose one variant or another from within their repertoire (or what a speaker’s repertoire might encompass at a given time). But the study of the relation between documented language and socio-historical settings is not fruitless. The fact that we have considerable documentation at our disposal for earlier and later Akkadian varieties is a valuable asset for analysing variation in the wider context of language change. In the case study concerning the representation of the phonemic unit /s/, diachrony plays a significant role in the distribution of variant spellings. Within Old Babylonian, the phonological context in which /s/ is embedded also predicts to a significant extent the different orthographic choices, strongly suggesting that the overall change in written texts ultimately responds to phonetic pressures on the spoken domain. However, the uneven adherence to this pattern in OB letters can also be positively associated with the geographical origin of the sender.

While the existence of different orthographic traditions ―rooted in scribal education, and independent of immediate spoken forms― can indeed explain the synchronic-diatopic variation, its mere existence is still highly informative for the modern researcher. Importantly, the distribution of more or less innovative /s/-variants in Old Babylonian centres does not necessarily align with innovation in other variables. That is, one area might exhibit a preference for the /s/-variant that later becomes widespread in Middle Babylonian, while simultaneously retaining more conservative features which are no longer kept in other areas. This suggests that variation in writing might be linked, at least in part, to linguistic changes spreading unevenly, with certain regions adopting certain innovations like the /s/-variant early, while still maintaining retentions that have in turn evolved differently in other regions.

The study of other traits can illustrate further aspects of variation in Old Babylonian. The distribution of the form šu’āti illustrates how analysing variation can help make sense of the links between the linguistic and socio-historical context of the time. Letters issued by (or on behalf of) King Hammurabi of Babylon consistently exhibit the pronominal form šu’āti (SG.ACC/GEN); at the same time, letters sent by the king’s right-hand man in administrative matters display a mixture of šu’āti and šâti, its contracted counterpart (Hernáiz, in preparation). The letters from both senders have identical recipients and are about the same topics, levelling out any apparent diaphasic distinction. However, the variation between these letters, all sent from Babylon’s central administration, becomes more relevant when an overview of the OB record shows the prominent use of šu’āti in Babylon’s area and a preference for contracted forms in southern regions ―precisely the place of origin of the king’s official. The study of variation and change in Old Babylonian, therefore, not only helps us better understand the diatopic aspect of the linguistic landscape of the time, but also reveals traits that testify to the different nuances with which social actors presented themselves in written documents. Whether there was a difference in how the king and his official used the pronominal form (or how their respective scribes did) cannot be ascertained; however, we can see that the letters issued on their behalf would contain distinctive traits that would represent the senders’ persona to the recipients of the missives. This is of interest for historical sociolinguistics as it relates to variation whereby a speaker’s activity appears embedded in varied and changing social contexts (Tuten; Tejedo-Herrero, 2011, p. 285–286). In cases like this, our knowledge about the individuals involved in the letter exchange is not far from complying with the requirements for sociolect delimitation expressed in Winter (1999, p. 75), whereby “text data have to be matched with findings about the text producers and users, about their specific group affiliations, and about their interaction among themselves and with others”. Further research on morphosyntactic variation, such as the presence or absence of gender syncretism in the pronominal paradigm (Hernáiz, in preparation) can expand the study of variation beyond the limits of what can be categorised as purely ‘spelling’ practices, delving into morphosyntactic variables.

The richer perspective of languages of the past revealed by analyses of variation (and change) exemplified in this article cannot be reconstructed on the basis of comparative linguistics (unless a clear signal in daughter or related languages can be detected). Language reconstructions give us crucial information about undocumented languages, inferred from later documented corpora, but they can hardly go beyond generalizations and regularities; the subtleties of language variation can easily fly under the radar. This article illustrates the presence of an intricate distribution of variants within Old Babylonian textual corpora, implying the likelihood of variation occurring at some point in the spoken domain. While this broad community of speakers included various Old Babylonian socio-political centres, there is evidence of continuous population movement suggesting that each centre would have their own spectrum of linguistic diversity (Efrat-Kowalsky, 2022). Consequently, the very specific aspect of the general uniformitarian principle that assumes that language must have been variable finds support in the written records of languages spoken as far back as 4000 years before the present. Further research might advance our assessment of other uniformitarian expectations, such as linguistic differences related to distinctive social groups and gender, or individual awareness of these differences (Bergs, 2012, p. 96). Even though the textual data and our knowledge about socio-linguistic structures for such distant temporal coordinates is decidedly limited, the Old Babylonian written record stands as a testament to the existence of variation in a time distant enough be paired with prehistorical communities and uniformly reconstructed proto-languages in other parts of the world.

Additional Information

Conflict of Interest

The author declares no competing interests.

Statement of Data Availability

The author confirms that the data and materials supporting the findings of this study are available within the article and its supplementary materials.

References

Adams, James N. (2013): Social variation and the Latin language. Cambridge University Press. doi:10.1017/CBO9780511843433.

Almansa-Villatoro, M. Victoria (2023): Variation as a social device: ‘Middle Egyptianisms’ in Old Kingdom letters. In Carlos Gracia Zamacona (ed.) Variability in the earlier Egyptian mortuary texts, 340–369. Brill. doi:10.1163/9789004677982_014.

Auer, Anita, Catharina Peersman, Simon Pickl, Gijsbert Rutten & Rik Vosters (2015): Historical sociolinguistics: the field and its future. Journal of historical sociolinguistics. De Gruyter Mouton. 1(1). 1–12. doi:10.1515/jhsl-2015-0001.

Bergs, Alexander (2012): The Uniformitarian principle and the risk of anachronisms in language and social history. In Juan Manuel Hernández-Campoy & Juan Camilo Conde-Silvestre (eds.) Handbook of historical sociolinguistics. John Wiley & Sons, Ltd. 80–98. doi:10.1002/9781118257227.ch5.

Carling, Gerd & Chundra Cathcart (2021): Reconstructing the evolution of Indo-European grammar. Language. Linguistic society of America 97(3). 561–598.

Charpin, Dominique (2010): Reading and writing in Babylon. Harvard University Press.

Charpin, Dominique, Dietz Otto Edzard & Marten Stol (2004): Mesopotamien: Die altbabylonische Zeit (Orbis Biblicus et Orientalis). Academic Press / Vandenhoeck & Ruprecht. 160/4 doi:10.5167/uzh-151595.

Clackson, James (2015): Language and society in the Greek and Roman worlds (Key Themes in ancient history). Cambridge University Press. doi:10.1017/CBO9781139047944.

Dunn, Michael, Simon J. Greenhill, Stephen C. Levinson & Russell D. Gray (2011): Evolved structure of language shows lineage-specific trends in word-order universals. Nature 473(7345). 79–82. doi:10.1038/nature09923.

Efrat-Kowalsky, Nour (2022): Text and the city: How the city shaped language. Old world: Journal of ancient Africa and Eurasia 2(1). 1–19. doi:10.1163/26670755-01010013.

Faber, Alice (1985): Akkadian evidence for proto-Semitic affricates. Journal of cuneiform studies 37(1). 101–107. doi:10.2307/1359962.

Garcia-Ventura, Agnès (2017): Emesal studies today: a preliminary assessment. In Lluís Feliu, Fumi Karahashi & Gonzalo Rubio (eds.) The first ninety years. A Sumerian celebration in honour of Miquel Civil, 145–158.

George, Andrew R. (2018): CUSAS 36: Old Babylonian texts in the Schøyen collection. Part one: Selected letters. Eisenbrauns.

Goetze, Albrecht (1945): The Akkadian dialects of the Old-Babylonian mathematical texts. In Otto Neugebauer & Abraham Sachs (eds.), Mathematical cuneiform texts, 146–151. American oriental society and American school of oriental research.

Goetze, Albrecht (1958): The Sibilants of Old Babylonian. Revue d’assyriologie et d’archéologie orientale 52(3). 137–149.

Hernáiz, Rodrigo (2020): Studies on linguistic and orthographic variation in Old Babylonian letters (Alter Orient und Altes Testament Band 465). Ugarit-Verlag.

Hernáiz, Rodrigo (in preparation): Paradigm lost: neutralisation of gender distinction and variation in Old Babylonian pronominals.

Huehnergard, John (2018): A grammar of Akkadian (Third Edition). Brill.

Jacquet, Antoine (2013): Family archives in Mesopotamia during the Old Babylonian period. In Michele Faraguna (eds.) Archives and archival documents in ancient societies: Legal documents in ancient societies IV: Trieste 30 September-1 October 2011 (Graeca Tergestina. Storia e Civiltà 1). EUT.

Jagersma, Bram (2010): A descriptive grammar of Sumerian. Doctoral thesis, Universiteit Leiden.

Kogan, Leonid (2011): Proto-semitic phonetics and phonology. In Stefan Weninger (ed.) The Semitic languages: An international handbook, 54–151. doi:10.1515/9783110251586.54.

Kouwenberg, N.J.C (2017): A Grammar of Old Assyrian. Brill.

Labov, William (1972): Sociolinguistic patterns (Conduct and communication series). University of Pennsylvania Press.

Lauersdorf, Mark R. (2021): Historical sociolinguistics and the necessity of interdisciplinary collaboration. In Allison Burkette & Tamara Warhol (eds.) Crossing borders, making connections (Interdisciplinary linguistics, vol 1), 207-230.

Mullen, Alex & Patrick James (eds.) (2012): Multilingualism in the Graeco-Roman worlds. Cambridge University Press. doi:10.1017/CBO9781139012775.

Nevalainen, Terttu & Helena Raumolin-Brunberg (2003): Historical sociolinguistics: Language change in Tudor and Stuart England. Longman.

Romaine, Suzanne (1988): Historical sociolinguistics: problems and methodology. In Ulrich Ammon,

Norbert Dittmar, Klaus J. Mattheier & Peter Trudgill (eds.), Sociolinguistics/Soziolinguistik: An international handbook of the science of language and society/ Ein internationales Handbuch zur Wissenschaft von Sprache und Gesellschaft, Volume 2, 1452-1468. De Gruyter.

Romaine, Suzanne (2009): Socio-historical linguistics: its status and methodology (Cambridge Studies in Linguistics 34). Cambridge University Press.

Rutten, Gijsbert & Marijke van der Wal (2014): Letters as loot: a sociolinguistic approach to seventeenth- and eighteenth-century Dutch (Advances in Historical Sociolinguistics volume 2). John Benjamins.

Sallaberger, Walther (1999): „Wenn Du mein Bruder bist,...“: Interaktion und Textgestaltung in altababylonischen Alltagsbriefen. Brill.

Soden, Wolfram von (1995): Grundriss der akkadischen Grammatik. Pontificio istituto biblico.

Sommerfeld, Walter (1995): Anm. §30. Grundriss der akkadischen Grammatik, 35–36. Pontificio istituto biblico.

Sommerfeld, Walter (2006): Varianten in der Keilschrift-Orthographie und die historische Phonologie des Akkadischen. In Gregorio del Olmo Lete, Lluís Feliu & Adelina Millet Albà (eds.), Šapal tibnim mû illakū: Studies presented to Joaquín Sanmartín on the occasion of his 65th birthday, 359–376. AUSA.

Sommerfeld, Walter (2021): Old Akkadian. In Juan Pablo Vita (ed.) History of the Akkadian language (2 vols), 513–663. Brill. doi:10.1163/9789004445215_012.

Song, Jae Jung (2018): Linguistic typology (Oxford textbooks in linguistics). Oxford University Press.

Streck, Michael P. (2006): Sibilants in the Old Babylonian texts of Hammurapi and of the governors in Qattunan. In Guy Deutscher & N. J. C. Kouwenberg (eds.) The Akkadian language in its Semitic context, 218-254.

Streck, Michael P. (2022): Old Babylonian grammar: Volume one. Brill.

Trudgill, Peter (2020a): Millennia of language change: Sociolinguistic studies in deep historical linguistics. Cambridge University Press. doi:10.1017/9781108769754.

Trudgill, Peter (2020b): Sociolinguistic typology and the uniformitarian hypothesis. In Mily Crevels & Pieter Muysken (eds.) Language dispersal, diversification, and contact. Oxford University Press. doi:10.1093/oso/9780198723813.003.0003.

Tuten, Donald N. & Fernando Tejedo-Herrero (2011): The relationship between historical linguistics and sociolinguistics. In Manuel Díaz Campos (ed.) The handbook of Hispanic sociolinguistics. 283–302. doi:10.1002/9781444393446.ch14.

Walkden, George (2019): The many faces of uniformitarianism in linguistics. Glossa: a journal of general linguistics 4(1) doi:10.5334/gjgl.888.

Westenholz, Aage (2006): Do not trust the assyriologists! Some remarks on the transliteration and normalization of Old Babylonian Akkadian. In Guy Deutscher & N.J.C. Kouwenberg (eds.) The Akkadian language in its Semitic context: studies in the Akkadian of the third and second millennium BC, 252–260.

Whiting, Robert M. (1987): Old Babylonian letters from Tell Asmar (Assyriological studies no. 22). Oriental Institute of the University of Chicago.

Winter, Werner (1999): Sociolinguistics and dead languages. In Ernst Håkon Jahr (ed.) Language change: Advances in historical sociolinguistics, 67–84. doi:10.1515/9783110807653.67.

Yakubovich, Ilya (2009): Sociolinguistics of the Luvian language. Brill.

Appendix

List of instances of consonant-vowel syllabograms rendering /s/ in Old Babylonian letters in (George 2018):

Figure 6.

Figure 7.

Figure 8.

Figure 9.

Review

DOI: https://doi.org/10.25189/2675-4916.2025.V6.N1.ID792.R

Editorial Decision

EDITOR 1: Joe Salmons

ORCID: https://orcid.org/

AFFILIATION: University of Wisconsin–Madison, Wisconsin, Estados Unidos.

-

EDITOR 2: Josh Brown

ORCID: https://orcid.org/

AFFILIATION: University of Wisconsin-Eau Claire, Wisconsin, Estados Unidos.

-

EDITOR 3: David Natvig

ORCID: https://orcid.org/

AFFILIATION: University of Wisconsin–Madison, Wisconsin, Estados Unidos.

-

DECISION LETTER: Much early work in historical sociolinguistics started with studies of European languages in the Early Modern period into the 19th century. This paper explores a greater time depth and a part of the world where we have less of this kind of work to date. Hernáiz documents change in sibilants in Old Babylonian letters, a type of sound change and a data source widely used in historical sociolinguistics. While current work increasingly goes farther back into history and covers settings around the world, this important contribution shows just how much we can learn about linguistic variation and change and how deep into the past we can go.

Rounds of Review

REVIEWER 1: Joe Salmons

ORCID: http://orcid.org/

AFFILIATION: University of Wisconsin–Madison, Wisconsin, Estados Unidos.

-

REVIEWER 2: Samuel Boyd

ORCID: https://orcid.org/

AFFILIATION: University of Colorado Boulder, Colorado, Brasil.

-

ROUND 1

REVIEWER 1

2024-07-04 | 02:58 PM

See attached documents. This is a nice paper.

-

REVIEWER 2

2024-09-04 | 07:32 PM

The article provides a very helpful and needed contribution in terms both of general linguistics and the data available for researchers in Old Akkadian as well as Assyriology and the theorization of variation, writing system, and speech. The author handles both primary and secondary literature on the subject well. The focus on regional variation in particular is very well done. My main areas of comment are:

1) There are a few grammatical issues in the prose that should be cleaned up (for example, "data" is a plural noun, so "these linguistic data are," not "this linguistic data is," which should instead be "datum").

2) The engagement with secondary literature could be better. In addition to the works cited, which are excellent and make the overall analysis compelling, see: Hasselbach on Sargonic Akkadian; I. J. Gelb's classic work and grammar on Old Akkadian; von Soden's reference grammar of Akkadian, which is better for such things than Huehnergard's intro grammar, though that is naturally excellent; though the focus is on Old Babylonian, the author might cite Kouwenberg on Old Assyrian; see also the edited volumes recently by Hasselbach-Andee as well as the two volumes edited by Juan Pablo-Vita. None of this would change the author's points, but perhaps update the already good bibliography a bit.

3) Again, not crucial, but the author might want to show some awareness that later in the history of Akkadian the script and pronunciation begin to break down, such that Late Babylonian letters indicate, compared to Aramaic duplicates, that the writing system no longer follows speech well. In other words, the same study may not be replicatable for the other end of the history of Akkadian.

These are minor comments on an otherwise excellent and publishable paper.

-

ROUND 2

REVIEWER 2

2024-12-10 | 01:48 PM

The revisions have all been made to my satisfaction. This is a wonderful article and I am happy to recommend it for publication.

Authors' Replay

DOI: https://doi.org/10.25189/2675-4916.2025.V6.N1.ID792.A

-

ROUND 1

2024-11-14

I would like to thank both reviewers for their excellent suggestions and their thoughtful feedback.

I have now eliminated some typos and mistakes (some of them, kindly spotted by the reviewers). I have renamed the last section (from “Discussion” to “Discussion: variation and change in Old Babylonian”). I have also reviewed the bibliography, and made some changes to maintain consistency in the reference entries.

My comments below (in red) are for the reviewers only.

Reviewer 1

This is a valuable piece of work and definitely should be included in the special issue. A few minor and relatively easy changes can make a strong paper stronger. In particular:

Nothing against Werner Winter (I actually had a class with him in Kiel in 1980!), but it’s hard to see his now very dated paper used as a frame for this very contemporary study. He surely couldn’t have imagined the explosion of corpus data or computer technical that makes a paper like this possible today. I urge you to reframe this in a more positive way: You and others are surmounting problems that were once insurmountable thanks to new methods, data and theories. It’s fine to mention Winter as an example of how things used to be and how we used to think, but we’re making real progress and you’re part of it.

This is an important issue, and I’m very glad that the reviewer has made me aware of it. I have reframed my wording (and deleted some less relevant mentions to the article) to present Winter’s comments in a more appropriate and fair manner, with special attention to stressing the time circumstances that differentiate our perspectives.

I think it’s important to your argument that this is a completely normal sound change, typologically found across many languages.

Noted and added now to the text.

Yes, everybody quotes Labov on ‘bad data’ but here too I urge you to take a positive mindset: Mark Lauersdorf implores us to ‘use all the data’ and argues that we can make lots of progress doing that. That’s exactly how you’re contributing here. (I’m happy to send Mark’s paper if you don’t know it.)

I have updated this, changing the perspective and substituting Labov’s reference for the interdisciplinary call made by Lauersdorf (2021).

The footnotes outlining the statistical analysis are important and should be moved to the body of the paper.

That’s a good idea. I have now added them to the main text.

I wouldn’t spend time and space on talking about how some other settings are a little older than this one … this is indeed the distant past. I might note older writing briefly, but nothing more.

I have now reduced this and presented it as a footnote, to keep the focus of the narration on Old Babylonian while still highlighting its predominance in terms of documentation and possibilities for research on variation, compared to other older languages.

Would it be possible to add a map so that we can easily see the key places discussed?

I have now added a map. If it cannot be a colour map, one can format the picture on Word to grayscale.

The comments at the very end of section 3 are really important. Would it be possible to highlight this point in the intro and conclusion?

I didn’t find a suitable place on the introduction, as it’s less focused on the case study, but I have now expanded these ideas in the final section.

I’m uploading a file with some typos and such noted. There’s also a little more minor stylistic work that can be done in later. (It’s usual to have an appendix rather than an annex, for example.)

Thanks ever so much!

Reviewer 2

The article provides a very helpful and needed contribution in terms both of general linguistics and the data available for researchers in Old Akkadian as well as Assyriology and the theorization of variation, writing system, and speech. The author handles both primary and secondary literature on the subject well. The focus on regional variation in particular is very well done. My main areas of comment are:

1) There are a few grammatical issues in the prose that should be cleaned up (for example, "data" is a plural noun, so "these linguistic data are," not "this linguistic data is," which should instead be "datum").

I have corrected some typos, including all references to ‘data’.

2) The engagement with secondary literature could be better. In addition to the works cited, which are excellent and make the overall analysis compelling, see: Hasselbach on Sargonic Akkadian; I. J. Gelb's classic work and grammar on Old Akkadian; von Soden's reference grammar of Akkadian, which is better for such things than Huehnergard's intro grammar, though that is naturally excellent; though the focus is on Old Babylonian, the author might cite Kouwenberg on Old Assyrian; see also the edited volumes recently by Hasselbach-Andee as well as the two volumes edited by Juan Pablo-Vita. None of this would change the author's points, but perhaps update the already good bibliography a bit.

I appreciate this suggestion. I have now added a reference to Kouwenberg’s grammar of Old Assyrian (even though Old Assyrian is only mentioned once). Von Soden’s grammar and one chapter of Vita are already cited in the article. I know well the other references, but I did not think it was necessary to cite them in the text. Sargonic Akkadian is hardly mentioned (even less prominently now, after its mention has been relegated to a footnote, after the suggestion from the other reviewer to devote less time to languages that are older than OB). Therefore, Hasselbach’s and Gelb’s grammars could only be added as expansions to footnotes, and I wonder whether that would be a bit too much?

3) Again, not crucial, but the author might want to show some awareness that later in the history of Akkadian the script and pronunciation begin to break down, such that Late Babylonian letters indicate, compared to Aramaic duplicates, that the writing system no longer follows speech well. In other words, the same study may not be replicatable for the other end of the history of Akkadian.

Good point. However, the article clearly focuses on Old Babylonian (+1000 years earlier than Late Babylonian), and I’m not sure it’s necessary to mention this. On the other hand, while the textual record and the linguistic ecology of Late Babylonian differs greatly from OB, I would prefer not to mention anything that could imply that a similar study on variation might not be fruitful, even if it’s for different parameters (such as the influence of Aramaic lects on written Akkadian).

These are minor comments on an otherwise excellent and publishable paper.

Thanks ever so much!