Special Supplemental Letter from the Editor SpecGram Vol CLIX, No ι Contents

Acquiring Isolation: The Peculiar Case of Ghwǘǜb

Tashel M. Kaithe and Valencia R. O’Shaughnessy
Montmorecy Institute of Cross-Disciplinary Studies

Ghwǘǜb, an Amazonian language spoken by a single community, initially appears to be the epitome of a linguistic isolate, perhaps the remnant of a formerly wide-spread language community whose speakers acquired the language of, or were displaced by, those of a different group moving into the area. Recent archeological evidence, as well as data from genetic survey research, has called this interpretation into question, however. This paper argues that an acquired isolation mechanism, driven by a confluence of genetic and cultural factors, is responsible for the “faux isolate” status of the language.

One of the more predictable byproducts of the ongoing enterprise among historical linguists to classify languages into families is the identification of isolates, languages which do not appear to have any extant relatives. The initial assessment of a language as an isolate is typically followed by intensive and careful examination of possible connections that may have been obscured via perfectly normal sound changes; this process has, over the years, decreased the number of languages considered true isolates, but some, of course, remain. Basque is the textbook example; despite varied attempts to relate it to other stocks, such as Etruscan, proto-Caucasian, or Sumerian, it remains unrelated to any other language.

Linguists widely agree on what the likely process is that gives rise to isolates: the replacement of surrounding related languages by those of a different family, leaving only one of the originals surviving. In a pure example, one might expect the speakers of this language to preserve cultural traditions different from those of the surrounding communities, but language and culture can be disjoined to some extent, and cultural influences can cross linguistic boundaries. One rarely encounters pure examples in real life. When one does run across an example that seems too perfect to be true, it is reasonable to be suspicious. And this brings us to Ghwǘǜb.

Ghwǘǜb is an Amazonian language, with one community of speakers located on the varzea of the Tairani River, a tributary of the Amazon located in the southwest of the central basin. Initial fieldwork, carried out during part of a survey of several tribal groups in the area, produced a picture of a people whose language (and lifeways) adhered so closely to the archetype of an isolate that the researchers felt it necessary to perform a more extensive study, which became even more strongly motivated when geneticists doing research in the same area announced results that ran exactly contrary to what would have been predicted by linguists, assuming even a very loose correlation between marriage patterns and language groupings. The following sections will detail the characteristics of the Ghwǘǜb-speaking community relevant to this discussion and propose that the etiology of these characteristics reveals a unique pattern of cultural motivations that can serve as a limiting case for historical linguistic predictions.

Context of this Study

Initial references to the Ghwǘǜb community are found in Schankelvekker (1898), who refers to them using the name Yaribogenadrea, a label given them by the neighboring Torikara. Schankelvekker did not encounter the Ghwǘǜb herself, and was doing only a preliminary anthropological survey of the Torikara, with whom she communicated via a Guaraní-based trade pidgin. ‘Yaribogenadrea’ is actually a Torikara descriptive, a nominalized participial form (cf. Schadenpoodle 1989):


Fortunately, the Torikara use this name only for the Ghwǘǜb, not any of the other tribes in the area, so matching them with Schankelvekker’s fieldwork was nonproblematic. Her notes describe the Ghwǘǜb and their language only briefly, as an aside to a discussion of the extensive relations the Torikara have with tribes in the area other than the Ghwǘǜb (p. 629; English translation by the authors):

Most of the villages along this stretch of the river trade regularly with one another, and there is a Torikara term, tasivaodo, [‘go-and-get-er’ —Ed.] for specific villagers who do this frequently; the trade networks also function for arranging marriages and the frequent potlatch-type dinners by which clusters of villages exchange surplus goods. The only tribe with which they do not often trade is the Yaribogenadrea, whose village lies to the southwest. The Torikara seem to find these people offensive in some way; much derisive laughter accompanied their description of the tribe. They occasionally do trade copper with the Yaribogenadrea, however, in return for the colorful feathers of a bird mostly indigenous to the latter’s territory. I was unable to visit their village, though, and therefore cannot include cranial measurements.1

No linguistic work on the Ghwǘǜb was done until that documented in Schadenpoodle (2008); the portions of this article dealing with linguistic structure are summarized from that monograph. Serendipitously, genetic survey work was being done in the area by Heeliks (2003) under the auspices of the Human Genome Endeavor (partly due to the discovery, reported in Thrennesson (2002), of a FOXP2 mutation and attendant language-learning impairment among some residents of the area), and Pinchtoddle’s cultural anthropological work with the Torikara led him to investigate the Ghwǘǜb as well in late 2006. The material below is compiled from the authors’ comparative notes.

Isolate vs. Non-Isolate

Status of Ghwǘǜb as a Linguistic Isolate

Ghwǘǜb is strikingly different from any of its close (or distant) neighbors, in its phonology and in its lexicon; indeed, it shares none of the areal features that otherwise are widely distributed across languages in this part of the basin. Its phonemic inventory, described in detail in Schadenpoodle (2008), is typologically unusual; it violates at least one putative universal by having more back vowels than front ones, and some of the affricates are particularly distinctive (please note that the authors, for sake of brevity, are adopting non-IPA symbols for several phonemes; the properties of these are discussed below):2

  Bilabial   Dental   Palatal   Velar 
Stops Voiceless p t   k
Voiced b d   g
Fricatives Voiceless φ   š  
Voiced β L ž γ
Affricates  Nasalized Ћ Ђ   Њ
Regular   λ    
Nasals   m n    
Liquids Unrounded   l    
Rounded   lw    
Glides Unaspirated      y w
Aspirated     yh wh
 Front   Central   Back 
High  Spread  i ı ɯ
Round ü   u
Nasal     ũ
Mid Spread e ə ɤ
Round ö   o
Nasal     õ
Low   æ

While it is conceivable, of course, that extensive historical change could have produced such a system from the same substrate which gave rise to languages like Torikara, it is difficult to lend credence to such an explanation; the number of changes needed to derive Ghwǘǜb phonology from any of the proto-forms posited for languages of this region (cf. Tanekis 2008) would suffice to derive Chinese from them as well, or Klingon for that matter.

Using the extended Swadesh list, and a match algorithm keyed to shared phonetic features but not phonemic identity (so, for example, both [t] and [n] were possible matches for [d]), Schadenpoodle (2008.452) carried out a lexical comparison of Ghwǘǜb with eight of the surrounding languages; he found only one word that matched that of another tribe: tarana ‘copper’the item they trade with the Torikara for. In contrast, Torikara shares over 60% of the list with Yakali and Nerduru, and the Kanarati/Qanradi/Wartaina cluster, which has an 85% internal overlap rate, also shares 35% with Torikara/Yakali/Nerduru. No overlap was found between Ghwǘǜb and more geographically distant Amazonian languages (e.g. Guaraní, Shipibo, Hixkaryana) as well.

The language likewise is unlike its neighbors syntactically, although given the wide variation in Amazonian languages in this regard such a result is perhaps less surprising. While word order in Ghwǘǜb is flexible, the most common, unmarked order is [O A S V], where A represents a required auxiliary-like lexeme:

whàka-žìmЋú    ö-waЊìЊ    glup-     tiš-nùzi-waáà
hut-m.aunt:Gen    Indic:Rep-Pres3    Glup-Sbj    chew.ann.-srf-rep-3h

‘I heard that Glup chewed annoyingly on his maternal aunt’s hut.’

Again, although this type of construction is dissimilar to that used by any other in the area, it is less convincing as an argument for isolate status than is the fact that none of the basic grammatical functors appear to be cognate to any lexeme in those languages.

Ghwǘǜb as a Cultural Isolate

While it is not uncommon for groups sharing the same geographic area to adopt distinctive cultural traits or artifacts as a way of preserving or enhancing group identity, the Ghwǘǜb seem to represent an extreme case. Four traits in particular, are shared by no other groups in the area (summarized from Pinchtoddle, 2006):

Domestication of the Sloth. Alone among Amazonian peoples, the Ghwǘǜb train sloths (three-toed) as a kind of living animal trap. Although the training process itself seems to take a very long time, the product is sloths that will fall from trees on animals passing underneath and then cling to themrather like an avalanche of adhesive bowling balls. They are also, apparently, trained to attack the victim, but the animal is usually recovered by hunters well before any effect of the sloths’ ferocity is detectable. One of the traditional Ghwǘǜb dances, done yearly over the course of four days, depicts a folk story of the origin of the practice, in which Gwiuiu, the Sloth God, offers to help Pləb, the first Ghwǘǜb; the ritual text has not been extensively translated as yet, but Gwiuiu seems to be motivated by a sense of cosmic ennui.3

The Potlatch Catapult. The chief, or bliaabip, of the village is expected to give potlatch-type feasts periodically (a trait which is shared with other groups in the region). Among the Ghwǘǜb, however, the chief launches food at individual community members using a device rather like a small catapult; catching the food in the mouth without use of the hands is considered a mark of social sophistication (the food item in question is a kind of ball, made up of sticky cooked taro paste with salted slugs embedded in it). The sole item of outside trade pursued by the tribe, copper, is in fact procured specifically for the construction of this device, which is not used in any other fashion. An individual’s order in the line leading to the catapult target site is determined by a complex set of factors.

A contemporary Ghwǘǜb bwàrp vessel (top diameter 72 cm.), used to hold taro paste during ceremonial feasts. The orientation of designs on Ghwǘǜb pottery is variable; the focal figure is at bottom right on this piece. The piece probably depicts folk-hero Plǝb charming slugs by using a nose-flute. The two figures above him are traditional representations of sloths. The wave-line design may indicate the underworld.
The yhoγЂuú. This is a type of musical instrument incorporating narrow tubes within which very thin slices of reed are positioned so as to produce high-pitched sounds when air moves over them. Distinctively, the instrument is inserted into the nose, rather than the mouth, and most players use slightly different ones in each nostril so as to produced variant effects. Adult male villagers usually play these as a group, although individual males play them as part of courtship rituals. Nose-flutes are not unknown among the world’s cultures, but they are not used elsewhere in this geographic region, and the Ghwǘǜb are rather distinctive in their deliberate manipulation of the instrument so as to produce sneezing at particular points. While Pinchtoddle has recorded performances (which can last up to six hours), they have not yet been analyzed by a musicologist. While it is a commonplace observation among anthropologists that music is not universalone has to be familiar with the conventions of a musical form to interpret itthe basic patterns and scales used in Ghwǘǜb performances are not used by surrounding groups at all. We played recordings to the Torikara, for example, who immediately judged them as unpleasant.4

Ceramics. The other groups in the area use pottery with patterns and motifs common to those associated with the expansion of Tupic groups (cf. Brochado and Lathrap 1982, Brochado 1984, Peixoto 1995), but Ghwǘǜb ceramics are distinctive (most containers have a trapezoidal cross-section, and decoration is primarily in the form of stick-figure abstractions of local wildlife, particularly sloths and slugs).5

Ghwǘǜb as a Genetic Non-Isolate

In complete contrast to evidence from linguistic and cultural patterns, DNA analyses consistently show that the Ghwǘǜb share a large number of markers that are otherwise uncommon in this region with neighboring groups, particularly the Torikara, Yakali, and Nerduru, forming what might be called the GTYN cluster (Heeliks 2003). Kanarati, Qanradi, and Wartaina share a subset of these markers as well, although otherwise show evidence of being closely related to Tupic groups, as might be expected. The GTYN cluster pattern emerges from both mitochondrial and Y-chromosomal analysis, indicating that relatedness is not simply a function of female outmarriage.

Furthermore, the distinctive FOXP2 variant first described by Thrennesson (2002) is also shared among members of the GTYN cluster, although only a few of those tested from the KQW cluster had it. Unlike the variant associated with SLI (cf. Lai et al. 2001; Gutiérrez-Clellen et al. 2009, Leonard 2009), this does not seem to cause difficulties with normal language use. Instead, it appears to cause greater difficulty with second language acquisition, although at present it is unclear whether this operates by intensification of L1 interference factors or (or additionally) by accelerating loss of neural plasticity after early youth. While the details of this “SLA impairment” disorder are still unclear, it appears to act similarly to a recessive trait; only learners who have both parents with the variant suffer from it. The tasivaodo among the Torikara, as might be expected, are individuals who do not suffer from the disorder, and Heeliks has noted that they tend to fall into familial units who preferentially marry tasivaodo from the Yakali and Nerduru.

It is important to note that the Ghwǘǜb community genotype itself is distinctive from the rest of the members of the cluster only in its homogeneitythere is more variation among other members of the GTYN cluster, and within the non-Ghwǘǜb communities in that cluster, than among members of the Ghwǘǜb. In short, the Ghwǘǜb do not, from a genetic standpoint, appear to be “outliers” at all, but rather occupy a median position to an unusual degree.

New Evidence, and a Possible Explanation

To this point, observations have strongly, almost overwhelmingly, supported a single interpretation of the Ghwǘǜb’s status as an isolate: Proto-Ghwǘǜb speakers inhabited the area prior to expansion of Proto-Tupic-speaking groups, and Proto-Tupic replaced Proto-Ghwǘǜb in some communities whose members were primarily from Proto-Ghwǘǜb-speaking groups. Thus, the GTYN cluster comprises modern descendants of Proto-Ghwǘǜb speakers, but only the Ghwǘǜb themselves continued to speak their mother tongue. The greater genetic homogeneity of the Ghwǘǜb may be ascribed to what might be termed anexogamy. Other cultural groups in this area practice systematic exogamyyoung women are expected to marry men from other groups. Endogamy, of course, is also known in other cultures, although it becomes difficult to maintain for groups under a particular size since continual reinforcement of recessive traits can be self-limiting in the long term. The Ghwǘǜb continuously attempt exogamic marriage, but are almost always unsuccessful. We could find evidence of only one case of a woman marrying into the Ghwǘǜb community; Yakisgo is elderly now, and has moved back to the Torikara, but as a young woman married a Ghwǘǜb man to cement trade relations. She has been deaf from birth, however, and her ability to provide information to researchers on short notice is limited, although her experience is, as we will argue below, crucial for understanding the historical status of the Ghwǘǜb.

Recently, Pinchtoddle’s work, along with archeological data from Arnheldsen’s (2007) survey of former village sites, has raised serious questions about the validity of the Ghwǘǜb-as-remnant account. Pinchtoddle noted that while the Qanradi and Wartaina tell extended traditional stories about how their ancestors traveled to their current location from a legendary homeland, and had to fight pre-existing locals for dominance, the enemies in the stories bear absolutely no resemblance to the Ghwǘǜb. Given the distinctiveness of the Ghwǘǜb, one would not expect storytellers to confuse them with other groups; at the very least, one would anticipate a mention of sloths, or even a modified version in which the autochthonous foes were allied with monsters of some sort. Nothing of the sort appears. The foes described in the story do, however, bear some resemblance to Tupic groupsthey are, for example, portrayed as having the secret of taro agriculture, which both the Qanradi and Wartaina say they “won” from these enemies. The process of transmuting history to myth is a notoriously variable one, and this counterargument to the Ghwǘǜb-as-autochthons position is therefore very tentative, but the pattern is intriguing nonetheless.

More concrete evidence comes from Arnheldsen’s observations of ceramic sequences. Test digs in four abandoned village sites, as well as in midden areas in three current villages (two Torikara, one Yakali), all found Tupic pottery below a layer in which typically distinctive Ghwǘǜb pottery shards occurred, with Tupic pottery again above that. Dating of the shards themselves, as well as other stratigraphic evidence, indicates that Ghwǘǜb pottery was brought into an area already inhabited by groups using Tupic motifs. Further supporting this sequence was the discovery at one abandoned village site (A435) of the remains of a small food-catapult, in a layer below one in which Tupic pottery was found, and above another.

In short, the Ghwǘǜb appear to be newcomers, not a remnant. This leaves us with an obvious question: Why is Ghwǘǜb a linguistic and cultural isolate? Its speakers came from somewhere, and their descendants still live in the area even though most of them don’t speak the language. But no language related to Ghwǘǜb has been found in any portion of South America (or elsewhere).

Yakisgo may provide an answer to the question, and raise new ones for linguistics at the same time. Anthropologists have long recognized the principles of linguistic and cultural relativity, but Ghwǘǜb may be a limiting case. We must note that the only apparently successful case of non-Ghwǘǜb marriage into the Ghwǘǜb community was deaf from birth. Alsoand we realize this claim will be controversial, to put it mildlyno non-Ghwǘǜb who hears the language is able to take the speakers seriously. Even Prof. Schadenpoodle, who not only has a well-deserved reputation for gravity but also has had rather traumatic encounters with the Ghwǘǜb, is incapable of hearing the language while maintaining the air of earnest respect any language deserves. Pinchtoddle’s (2006) fieldwork showed that even among the Ghwǘǜb, the relationship between language use and maintenance of social and reproductive relations appears vexed. In other cultures, language used in courtship rituals is geared toward enhancing the speaker’s or hearer’s statusto show one is a worthy mate, or to flatter the object of desire. In the Ghwǘǜb community, on the other hand, it is commonly acknowledged that courtship language is successful (or not) only because the hearer knows that if she capitulates, the speaker will stop talking.6

Basically, it appears that the Ghwǘǜb difficulty with outmarriage is a direct result of their language, and we surmise that most Proto-Ghwǘǜb speakers, after encountering speakers of Proto-Tupi, immediately and enthusiastically abandoned their mother tongue, along with many of their own cultural practices. However, some Proto-Ghwǘǜb speakers had inherited the FOXP2 variant limiting their SLA ability, and hence could not assimilate to Proto-Tupic. If we assume that the variant was originally found within a particular group of Proto-Ghwǘǜb speakers (i.e., that SLA-impaired speakers were already associated into a particular tribal unit, or band), we can thus account for the formation of the current situation. Having over the course of several generations become cultural outsiders, these speakers’ status as a separate people was continually reinforced. Normally, some of their offspring would lack the disorder (since it acts like a recessive trait), and the community would slowly dwindle as outmarriage thinned its ranks, but the social isolation of the community, together with the decidedly anaphrodesiac effects of being a Ghwǘǜb-speaker, encourage endogamy, whether it is voluntary or not. Thus, the variant was reinforced to the point where the entire community possesses it.


Limiting cases such as the Ghwǘǜb are vitalthey illustrate the strengths of a position by articulating its opposite and drawing attention to the opposite’s rarity. Linguistic relativity is not a principle adopted by linguists only because we have trouble measuring the “value” of languages, or because we have seen what horrors cultural bigotry can bring forth. Before Ghwǘǜb, linguists had no examples to analyze of a language that fundamentally doesn’t work wellany that may have existed in the past would simply have gone extinct as soon as speakers encountered a language that did work well. Modern cases of language death are not analogous; speakers of Cornish, for example, didn’t learn English because they immediately thought it sounded much better, but rather because they thought that learning English decreased the chance that the English would kill them, or at least might make the teacher stop hitting them; similarly, the reasons for Italic languages other than Latin dying out weren’t just legion, but also legions.

We are, of course, keenly interested in the question of how Proto-Ghwǘǜb developed. Its existence today reminds us that potentially a very, very large number of language families have existed at some point in the past but left no trace of their existence. Were it not for a genetic accident, Ghwǘǜb would have vanished as well, and as linguists, we can only be thankful that the language, like the sloths of its speakers, approaches the horizon only slowly.

Works Cited

Arnheldsen, Heraklios (2007). An initial survey of village sites in the Tairani headwaters region. Amazonian Archeological Bulletin 32(1):15-86.

Brochado, J.; and Lathrap, D. (1982). Chronologies in the New World: Amazonia. Unpublished manuscript. On file at the Department of Anthropology, University of Illinois at Urbana-Champaign. Illinois.

Brochado, J. (1984). An Ecological Model of the Spread of Pottery and Agriculture into Eastern South America. PhD dissertation, University of Illinois, Urbana. University Microfilms, Ann Arbor.

Gutiérrez-Clellen, Vera; Simon-Cereijido, Gabriela; and Erickson Leone, Angela (2009). Code-switching in children with specific language impairment. International Journal of Bilingualism 13(1): 91-109.

Heeliks, Jacob (2003). A genotypic cluster analysis of fifteen upper Tairani communities. Papers in Amazonian Genetics 4(2).

Lai, C. S. L.; Fisher, S. E.; Hurst, J. A.; Vargha-Khadem, F.; and Monaco, A. P. (2001). A forkhead-domain gene is mutated in a severe speech and language disorder. Nature 413(6855): 519-23.

Leonard, Laurence (2009). Some reflections on the study of children with specific language impairment. Child Language Teaching and Therapy 25(2): 169-171.

Peixoto, Jose Luis (1995). A ocupação do Tupiguarani na borda oeste do pantanal Sul-matogrossense: maciço do Urucu. M.A. thesis. IFCH-PUCRS, Porto Alegre.

Pinchtoddle, Vasily (2006). The Repose of Travel: An Ethnography of the Torikara. Scranton, PA: Psammeticus Press.

Schadenpoodle, Athanasious (1989). An initial sketch of Torikara syntax. Occasional Papers in Amazonian Linguistics 12.

Schadenpoodle, Athanasious (2008). A Grammar of the Ghwǘǜb Language. Scranton, PA: Psammeticus Press.

Schadenpoodle, Athanasious (forthcoming). Toward the resolution of certain problems in the phonemic analysis of Ghwǘǜb.

Schankelvekker, Hortense (1898). Eine vorläufige Beschreibung der Völker der Obertairaniregion, mit einem Nachtrag über Kranialmessung. Monograph published as separate issue of the Bulletin der Archäologischen Gesellschaft von Schleswig-Holstein 15.

Tanekis, Akioni (2008). Divergence models for Ghwǘǜb are less probable than the null hypothesis. Working Papers in Statistical Diachronic Linguistics 18(3):27-34.

Thrennesson, Olaf (2002). A novel FOXP2 variant is linked to language-learning impairment. Journal of Genetics and Bioinformatics 12(4):127-60.

1 Like many an anthropologist of her era, Schankelvekker was primarily interested in carefully measuring the cranial shapes of those she studied (the Torikara apparently passed down stories of her and her vigorous use of calipers in the form of their term for anthropologistsdagemaoda ‘crab-pincer-ghost’, as opposed to their usual term for Europeans, kedagwamaoda ‘bad-dancing ghosts’).

2 We realize that that the degree to which this phonemic inventory is unusual will likely be taken by the reader as evidence that the analysis on which it is based is probably flaweda quite understandable reaction. More work on the language is obviously badly needed, and the authors would be quite happy to find that there is a perfectly normal phonemic system underlying this apparently highly deviant one. However, we have independently found minimal pairs for the distinctions described, so if there is a problem, it is not one resulting simply from a problem with Schadenpoodle’s field methodology.

3 Schadenpoodle (pc) has attested in great detail that such sloths are quite capable of immobilizing even a fully-grown adult human.

4 It should be noted at this point that the Torikara are not in the habit otherwise of being highly judgmental about musicthey prefer their own, of course, but according to Pinchtoddle, they react to most unfamiliar music forms with a kind of tolerant, muted curiosity.

5 The slug figures are particularly distinctive, since no other extant culture, to our knowledge, has ever applied stick-figure abstraction to slugs, or even any other mollusk, land-bound or otherwise.

6 The phrase used in Ghwǘǜb to refer to a successful courtship translates literally as “She got him to shut up”.

Special Supplemental Letter from the Editor
SpecGram Vol CLIX, No ι Contents