Blog | Benjamin Molineaux (Edinburgh)

Heritage texts, typology and human history in the Southern Cone

Sat, 06 Sep 2025 11:00:00 +0000

In July 2025, I was awarded a 4-year Leverhulme Research Grant for the project “Texts, typology, and language history in the Southern Cone: A digital framework”. The research, focusing on language filiation, contact and change in some 30 languages, will be conducted by an 8-person team in Scotland, Chile and Argentina.

The Americas were the last continental mass to be populated by humans, who would have crossed via the Bering Strait and migrated southward following the last Ice Age, reaching Tierra del Fuego quickly thereafter. Despite this, genetic, archaeological and linguistic data seem to suggest that populations of the Southern Cone remained fairly isolated. In linguistic terms, this is reflected by the lower density of languages, and the absence of large families, as compared, for instance, to the Amazon and Gran Chaco zones immediately to the north. In fact, current descriptions of the Patagonia and Tierra del Fuego tend to find only smaller language families (like Chon or Huarpean) alongside numerous isolates – languages with no known relatives. Such descriptions, however, are based on minimal, poorly contextualised data. Our team’s efforts to gather and carefully annotate the linguistic materials should allow for a re-assessment of these claims with the best data available, establishing shared words, sounds and grammatical patterns which may point to links within and outside the area, thus probing isolationist theories.

Crucially, the project will bring together archival materials which have been difficult to access heretofore, creating a digital repository with abundant linguistic and contextual annotations. This resource, the Comparative-Historical Corpus of the Southern Cone (CHiCo-SC), will be available to researches as well as native communities, most of which no longer speak their heritage language.

While much of the early textual material from the Americas is problematic due to the interference of non-native writers, the prevalence of scholastic grammars, genre limitations, etc., there is no doubt that such materials represent versions of the Native American languages which must be squared with reconstructions based on current varieties.

Mapudungun phonology and morphology

Wed, 06 Feb 2019 11:00:00 +0000

In my doctoral research I investigated the synchronic perception and diachronic development of stress in Mapudungun ([ma.pu.θu.ˈŋun] – formerly ‘Araucanian’), a presumed language isolate spoken by the Mapuche ethnic people of south-central Chile and Argentina. The language’s stress-system is under-explored and, as an agglutinating, weight-sensitive language, provides an interesting view into the prosody-morphology relationship. I also believe that scholarly attention can lead to greater social acceptance of this marginalised language, promoting it as a valid means of education and participation in Chilean and Argentinian society.

I was also granted a to work on a corpus-based historical phonology and morphology of the language. The first version of the was the result of this project.

Below: Principal Mapudungun-speaking areas today (based on Adelaar and Muysken, 2004: 503):

Native American Philology

Wed, 06 Feb 2019 11:00:00 +0000

In 2021, I launched the first version of the which uses text-based approaches to reconstruct the 400-year history of this Native American language. Such tools, though mostly neglected for historical material of the Americas, have huge potential for pushing back the date of comparative reconstruction, informing historical developments of individual varieties, as well as ideas of local contact and genetic inheritance.

Languages of the Americas have, at best, only five full centuries of attestations using alphabetic writing. However, for some languages — especially for those deemed by missionaries to be ‘General Languages’ for the region — this data is abundant and well suited to exploring changes. Such a timespan, though more restricted than that which we find for many Old World languages, is nevertheless substantial enough to observe non-trivial changes in languages, especially where social pressures and population dynamics are rapidly changing. A quick look at colonial varieties of European languages — which share this timescale — should corroborate these facts.

The CHM — and future resources I am developing for the Southern Cone as part of the CHiCo-SC (incl. Chonan, Huarpean, Alacalufan, Yaghan and Kunza) — will provide searchable analyses of the lexical, morpho-syntactic, phonological and orthographical structure, proposing changes to account for variations across time, space and text.

A page from the earliest set of Mapudungun sermons, by Spanish Jesuit Fr. Luys de Valdivia, 1621.

Sound and spelling variation in Older Scots

Tue, 06 Feb 2018 11:00:00 +0000

I am a team-member in the AHRC-funded project, mapping sounds to spellings in the earliest documentary records for the Scots Language. Our source materials are the etymologically Germanic root morphemes in the tagged texts of the Linguistic Atlas of Older Scots ( ).

Our objective is to try to track individual morphemes and segments’ development through time and space, in order to (re)build the phonological history of the earliest period of the language “from the bottom up”. To this effect, spelling – and variation therein – is crucial. We will be mapping written forms to presumed sounds, and placing their variation across time and space in order to retrieve representative patterns.

A sample mapping of sounds to spellings (and vice-versa), from the FITS database:

The diachronic interaction of prosodic structure and morphology

Mon, 06 Feb 2017 11:00:00 +0000

To communicate, human languages must resolve the issue of grouping units of meaning by using elements that are predominantly formal, and which themselves are grouped into larger formal units. My interest, overall, is in the interaction of these two types of units, those of meaning (morphology) and those that constitute larger formal domains (prosody). Since parsing of prosodic and morphological structure is not always isomorphic, the two systems need to find common ground, one level of structure acquiescing to the other’s requirements.

As a result, some languages will display features that are morphology-strong, while others will be prosody-strong. A key fact determining these options will be the language’s morphological type, with some requiring more transparency in the morphology (agglutination, polysynthesis) and others less (fusion, isolation).