REMARKS ON THE VOCABULARIES.
BY R. G. LATHAM, M.D.
The observation which is the most necessary to the general student, as a preliminary to Mr. Wallace’s tables, is the nature of certain syllables which in the Uainambeu, Juri, Tariana, and other lists, appear at the beginning of the names of the different parts of the human body. It is certainly not by accident that in one language they all begin witheri-, in another withn-; and so on. It is equally certain that these prefixes are no part of the original word. In the Baniwa of the Rio Isanna, the term forhandiscapi; the syllablenu-is anon-radical prefix. Now the non-radical prefix is (almost to a certainty) the possessive pronoun, so thatnu-cabi =my head: the amalgamation of the two elements being common in the American, as well as in certain languages of the Old World. (Appendix to Macgillivray’s ‘Voyage of the Rattlesnake’—Languages of the Louisiade.) The distribution of these prefixed possessive pronouns is a better guide than theaffinities of the substantives themselves: since the personal pronouns are parts of speech which languages are slow to borrow from each other.
To apply the observation to the details of the vocabularies under notice, we find that the Juri and Javita prefixes are different from each other, and different from those of the Tariana, etc. throughout,—the pronominal syllables beingtch, andwa, respectively. Neither of these elements lead to anything very clear and patent in the way of affinity. On the other hand it is equally clear that in the Tariana, Isanna, Barré, and Tomo-Maroa lists the prefix (-n-) is constant, however much the roots which follow it may differ; so that the inference, in favour of the possessive pronouns (at least) being the same throughout those four tongues, is legitimate. In the Uainambeu the case is slightly different. It is only in some of the words that the prefix is-n: in others it is-eri. This however is not very material, since the two forms, in all probability, represent two persons,—nu=my, anderi=thy(orvice versâ) respectively.
Without, then, taking cognizance of the roots at all, the classification of the languages before us, according to the similarity of their pronominal prefixes, is as follows:—
a. Allied.—Uainambeu, Tariana, Isanna, Barré, Tomo-Maroa.
b. Disconnected.—Juri, Javíta.
c. Uncertain.—Lingoa Geral, Coretu, Cobeu, Tucano.
The geographical localities of these tongues coincide with the nature of their pronominal prefix, and favour (pro tanto) the notion that they all belong to one and the same class,—a class of which the value is at present wholly conjectural.
How far are the roots themselves similar? Upon the whole it may be said that where the geography and where the pronominal prefixes indicate affinity, the roots themselves do the same; though not so clearly and patently as the investigator unpractised in American philology has a tendency to expect.To go into the reasons of this would take up much time and paper. It is sufficient to state that, whether the percentage of similar words be great or small, it tallies with the similarity of the prefixes. This may be verified by noticing the distribution of such words as the following:—
Head= rootb-din Uainambeu, Tariana (?), Isanna (?).
Mouth= rootn-min Uainambeu, Tariana, Isanna, Barré, Tomo-Maroa.
Eye= root-tor-d, in Uainambeu, Tariana, Isanna, Barré.
Nose= roott-kin Uainambeu, Tariana, Isanna.
Teeth= root-i-in Uainambeu, Tariana, Isanna, Barré.
This list may be extended, but the foregoing words suffice for illustration. That the root, in many cases, agrees where the pronouns differ is evident: but it must be remembered that the position to which the present writer commits himself is simply that of thegreatest amount of radical affinities going along with the greatest amount of pronominal similarity. He by no means asserts, that where the pronouns differ everything else differs also.
When specimens of a language are laid before the public, and such specimens are, at one and the same time, limited in extent and the first of their kind, it is rarely safe to go beyond the indication of their probable affinities, and a general sketch of the class they illustrate. This is the reason why the present writer limits himself to observations of the miscellaneous and unsystematic character of those here made.
The Barré.—The Barré forms the centre of a group—a group of which the value is uncertain. The other members of this, are the Baniwa of the Tomo and Maroa to the north, the Uainambeu to the south, the Tariana, and Baniwa of the Isanna to the west.
This statement (as has been already suggested) lies in the identity of the prefixn-, throughout.
The name requires notice. In Humboldt’s account of the population between the Orinooko and Amazon, we meet witha notice of thePoignavi, called also Gui-punavi: which are probably neither more nor less thanBaniwa, as modified by the pronunciation of a different dialect or language. The localities coincide nearly as closely as the name. That branches however of the same population are denoted by these forms is by no means certain; inasmuch as the name may be applied by one population to several,—just as the rootWel-sh (a term equally foreign to Wales, Italy, Wallachia, and the French portion of Belgium, as a native name) is applied to the Germans, to the Welsh, the Italians, the Wallachians, and the Walloons.
The Poignavi of Humboldt are said to speak a form of the Maypure; which is likely enough, and which is by no means incompatible with their affinities to the Baniwa. A full vocabulary however of their language is wanting. In place hereof we have but two words,—
Of theseoueniis Baniwa.
Further notice of the name will occur when we come to the Baniwa of Javita. That theManivasof the maps areBaniwais suggested by Mr. Wallace; and probably the E-quinabiare but little different.
The Juri.—The geographical relations of the Juri to the other languages of Mr. Wallace’s list may be seen on the map. The population which uses it lies south of the Japura, and probably on the water-system (we can scarcely apply the wordvalleyto these vast levels) of the Iça or Putumayo. It is perhaps as far from the Uainambeu, as the Uainambeu is from the Isanna; and certainly farther from any other member of the class last under notice than any such members are from each other. Hence, theprimâ facieview afforded by its geographical position is in favour of comparative isolation, at leastas far as regards the tongues which delight in the use of then-prefix.
Neither is it in any very close geographical contact with the Tucano and Coretu, the nearest of the other languages.
It is comparativelyisolated, as far as the languages of the present tables are concerned.
Now if we go from these to the ordinary maps, we find the contiguous populations to bear the names Tapaxana, Cambeva, Ticuna, etc., etc. What are the relations of the Juri tongue in this direction? Unknown. We have no specimens of the Ticuna, no specimens of the Cambeva, no specimens of the Tapaxana. There may be anything or nothing in the way of likeness; anything or nothing in the way of difference.
The Baniwa of Javita.—The relation of this to the other Baniwas must be determined by the vocabulary itself: since (as has been already suggested) the identity of name goes for nothing either way. It proves nothing in favour of affinity; nothing in favour of difference.
The pronominal prefixes (wa-) are different; but this again is onlyprimâ facieevidence of real difference. A pronoun of a different person (as has also been already suggested) may have been used.
The Lingoa Geral.—Last in the order of notice, though first in the list, is the Lingoa Geral. The basis of this is the Guarani language; and, with two exceptions, the distribution of the numerous dialects and subdialects of the Guarani tongue is the most remarkable in the world; the exception lying with the Malay and the Athabascan tongues. The Malay numerals have long been known to extend from Polynesia on one side to Madagascar on the other, and, along with the numerals, a notable percentage of other words as well. The Athabascan dialects are spoken on the shores of the Arctic Sea, at the mouth of the River Columbia, and within the tropics; since the speech of certain of the formidable Apach tribes of Mexico has been shown to belong to the same classwith that of the Loucheux at the month of the Mackenzie River.
These are the two most remarkable instances of a wide and irregular distribution of language known to investigators; and next to these comes, as already stated, that of the Guarani tongues.
It matters little from what point we begin to consider them. Perhaps the mouth of the Amazons is as convenient as any. If this be our starting-point, we may follow the coast southwards, and in the direction of the River Plate. In nine cases out of ten, as often as the earlier Portuguese adventurers came upon an Indian population occupying the sea-shore, that population spoke a language which they calledTupi,Tupinaki,Tupinambi, or something similar in the way of a compound of the roottup, and which they found to be intelligible with the forms of speech spoken in several distant districts elsewhere. If they landed on the parts about Bahia, the language was akin to what they had previously heard at Olinda, and what they would afterwards hear at Rio Janeiro: and so on along the whole sea-board. Hence, there were Tupi forms of speech as far north as the Island of Marajo, Tupi forms of speech as far south as Monte Video, and Tupi forms of speech in all (or nearly all) the intervening points of coast. The fishermen of the Laguna de los Patos spoke Tupi. The Cahetes of Bahia did the same. So did the Tamoyos of the Bay of Rio Janeiro, and so the Tupinaki, Tupinambi, and Tupinaes—the Tupi Proper so to say. This made the Tupi pass for the leading language of Brazil; so long at least as Brazil was known imperfectly, or along the sea-coast only.
It was soon however noticed that, as a general rule, the Tupi was spoken to only an inconsiderable distance inland,i.e.until one got to the Province of San Paolo, going southwards. In Goyaz, in the hill-ranges of Pernambuco, Bahia, Porto Seguro, etc., came forms of speech which those who spoke the Tupi separated from their own,—forms of speech of thebarbarians (so to say) of the interior as opposed to the more civilized mariners of the coast. The Botocudo, the Canarin, the Coroado, the Coropo, the Machacari, the Camacan, the Penhami, the Kirivi, the Sabuja, the Gran, the Timbyra, and a vast list of other Brazilian Indians besides, were different from and other than the Tupi. But this distinction between the coastmen and the inlanders ceases as we go southwards; and when Cabeza de Vaca, in 1540, made his overland journey fromSt.Catalina to the city of Assumcio on the Parana, the chief Indians with whom he came in contact were allied to each other and allied to the tribes of the coast.
But the nomenclature changed with the change from Portuguese to Spanish dominion; and, though the Indians of the Brazilian province of San Paolo (the province of Brazil where they first began to extend inland and towards the centre of the continent) and those of Rio Grande de Sul might be Tupi, the allied populations of Entre Rios, Corrientes, Monte Video, and Paraguay were known under the designation ofGuarani. This gave us aTupi-Guaraniclass of languages in which it was not very incorrect to say that the Tupi were the Guarani of Brazil, and the Guarani the Tupi of Paraguay. Thechiefdifference was a verbal and nominal one.
The first part, then, of the South American continent where the Tupi-Guarani (or Guarani-Tupi) were found in large masses, with an extension inland as well as an extension coastwise, were the Brazilian provinces of San Paolo in its southern part, the Brazilian province of Rio Grande de Sul, and the Spanish territories of Entre Rios, Corrientes, part of Buenos Ayres, Monte Video, and Paraguay—Paraguay in its southern and eastern rather than in its northern and western parts. In these quarters the language changed, and the affinities of the populations were with the Indians of the Chaco—the Mocobi, Toba, and Abiponians, etc.
It was through the Guarani of Paraguay that the Jesuits did the chief part of their labours in the way of conversion,and it was through the Tupi of Brazil that the Portuguese did the chief part of their Indian trading: facts which favoured the formation of a Lingua Franca on a Tupi-Guarani basis.
At present we have seen the Indians of this family following the coast or the water-system of the Plata. But there is another series of fact connected with their distribution, clear and patent enough to have been known to the earlier philologists, such as Hervas and his contemporaries. The extension of populations akin to the Tupi-Guarani was just as great along the course of the Amazon as it was along that of the Parana,—just as great, and probably greater. It was also spread along the feeders of the Amazon; the southern feeders, and the northern feeders of that river. By the southern feeders the line ran towards the system of the Parana, by the northern towards that of the Orinoco. So that as far northwards as the equator, and as far south as Buenos Ayres, Tupi-Guarani were to be found. More than this, as the Amazon arises on the Andes, and as its head-waters were Tupi-Guarani, the extension of that family from east to west was as remarkable as its diffusion from south to north. Still it gives us the hem of the garment, the fringe of the carpet, rather than the carpet or the garment itself. It gave the circumference of a circle rather than the parts pertaining to the centre. I call it a circle for the sake of illustration; for the sake of showing that there were a vast class of distinct and different languages which it had more or less completely surrounded. In reality the outline was irregular, formed in the main by the Ocean and the Amazons, but with long processes attached to it; dipping in towards the centre on one side, and flying off from the periphery on another.
We will go further in the details of the same processes, offsets, indentations, or prolongations. The Omaguas are found as far west as the Napo; and they preponderate at the junction of the Japura. They are said to speak the Yete,Putumayo, and Zeokeyo dialects of the Sucumbia language. This gives us a third term; possibly a third philological division—Tupi, Guarani, Sucumbia. Such are the names of the languages, the populations being Tupi, Guarani, and Omagua; the Tupi (chiefly) in Brazil, the Guarani in Paraguay, the Omagua in New Granada. The full details of the Tupi-Guaranis of the Amazons are somewhat obscure. At the mouth of the Tapajos, and the mouth of the Tocantins, they are specially mentioned, so they are in the great island of the Tupenambazes.
The Napo and Putumayo are the chief rivers that carry us along the Tupi-Guarani (Omagua) lines northwards. The most important of the southern lines are those of the Ucayale and the Huallaga, rivers, be it observed, of the extreme west; the latter running at the very foot of the Andes. The Cocamas and the Cocamillas occupy the watershed of these two streams for an undetermined distance southwards; and they navigate them with the boldness of the Omaguas. This gives us a fourth section—Tupi, Guarani, Omagua, Cocamilla.
Fifthly and (in the present state of our knowledge) lastly, come the Indians of the Peruvian province of Santa Cruz de la Sierra, extended, on one side, into the Republic of Tarija and the Chaco, and, on another, into the Mission of Chiquitos. Three sections of the division have been visited and described—a, the Guarayos;b, the Chiriguanos;c, and the Sirionos. These are described by D’Orbigny in his ‘Homme Américain,’ pp. 338–348.
It is submitted that, in the question as to the originalhabitatandfocusof the Guarani family theprimâ facieevidence is in favour of the Provinces of Santa Cruz de la Sierra; the locality of the Guarayos, Chiriguanos and Sirionos, the two latter bearing, probably, the same name in a modified form. This view is proposed because the locality in question is the only portion of the Guarani occupancy which is common to the two great water-systems on which they are soremarkably distributed; and it rests upon the principle of not multiplying causes unnecessarily, rather than upon any special evidence. In one respect indeed, there is an accredited fact against it. The Chiriguanos are said to have been pressed forward into their present locality by the Spanish conquest—a fact, however, of which the evidence is very unsatisfactory, and one which does not apply to the Guarayos. I repeat the statement that the present notice goes no further than the indication of theprimâ facieprobabilities of the question, and that it merely points out the most promising line of criticism, a line which has yet to be worked out.
But there is a point of detail which may well find its place here. The languages which have hitherto been placed in the Guarani class are languages of which the Guarani character has long been admitted. Of new additions no notice has been taken; yet one such new addition can probably be made—one, and perhaps more.
Into the details of a language spoken to the south of the Amazons, and to the east of the Madeira, the language of the Mundrucu Indians, I have gone no further than the short list of Balbi’s ‘Atlas Ethnographique.’ However, the inspection of this is sufficient to show that it is more Guarani than aught else.—e. g.
If this view be tenable, theuaschiof Mundrucu simply =light, and its difference in form from the Guarani words is explained. So is its similarity to the nameuashiat=moon.
The Coretu.—It by no means follows that because two languages have the same name they are identical; on the contrary, they may be widely different—a point which has already been noticed. Now the ‘Atlas Ethnologique’ of Balbi gives us a short Coretu vocabulary which isnotthe Coretu of Mr. Wallace.
In these two lists the only two similar words are the names forsun. Hence, the phenomenon illustrated by the nameBaniwais repeated in the case of the wordCoretu.
Upon theCobeuandTucano, no facts beyond those that lie in the tables themselves are known.
With the exception of the Lingoa Geral, Mr. Wallace’s vocabularies represent languages hitherto unknown. What is their geographical relation to the known ones?
It is only through the northern tongues (the Baniwa of Javita and the Tomo) that we even approach the areas of anywell-described dialect. The Uainambeu in the south-west, the Juri in the south, and the Coretu in the west, are each and all on the limits ofterræ incognitæ. For the parts between the western watershed of the Rio Magdalena in New Granada, the 71st degree of west longitude, the 4th degree of north latitude (there or thereabouts), the River Napo, and the Amazons, I know of no vocabularies, still less of any grammars. Hence, of any tongue spoken at one and the same time to the west of the Rio Negro, to the south of the Rio Inirida, to the north of the Amazons, and to the east of the Putumayo, the only specimens are the ones under notice.
1. To the east of the Rio Negro, the vocabularies that bring us nearest to Mr. Wallace’s are those of Sir Robert Schomburgk, of which the Guinau is the most western. They are all dealt with by Sir Robert Schomburgk himself as members of the great Carib (Carib-Tamanak) family, and this upon reasonable and sufficient grounds.
2. For the north we must seek our chief data in the ‘Mithridates’ and in Humboldt.
Along the rivers Meta, Vichada, and Guaviare, feeders of the Orinoco, different forms of the Saliva language are spoken—a language of which the distribution reminds us of the Guarani, although it is far less remarkable for its extent. At the same time it is so farfluviatileas to follow to the system of the Orinoco, and so far extensive as to occur on the Upper Meta at the present time, and to be supposed (on reasonable grounds) to have once reached as far eastwards as Trinidad.
b.Conterminous with the Saliva, and also conterminous with the Guiana and Venezuelan members of the great Carib family, lie the populations speaking languages akin to the Maypure and Pareni—of both of which forms of speech we have specimens (Humboldt’s and that of the ‘Mithridates’), though short and insufficient. That the Caveri, the Avani, and the Poignavi (Guipunavi) speak dialects akin to eachother and to the Maypure rests upon external evidence. Specimens are wanted. South of the Rio Uapes in Arrowsmith’s London Atlas lie the Meppuris, similar in name to the Maypure, but by no means necessarily allied to them.
c.Further north (in a north-western direction) on the Casanare and the Lower Meta, are the Yarura, Betoi, and Otomaca tongues.
3. South of the Amazons we must descend as far as the Province of Moxos (on the Beni) before we get anything beyond the most fragmentary specimens of language.
4.Westwards, and in the direction of the Andes, the break is greater still. For New Granada, a few words of the old Muysca language from the parts about Tunja, and, then, a short list from the mouth of the Atrato (at the very neck of the Isthmus of Darien) constitute the whole of our materials.
b.And matters are but little better in Ecuador. Between the Andes (of which the different Quichua dialects are pretty well known) and the area now under notice, the Zapara vocabulary of Osculati’s recently published work is all we have. The Zapara is spoken on the Rio Napo.
Hence, if we pass from the area illustrated by the specimens before us—specimens from aterra incognita—to the region of known dialects, we have (after a few Puru and Mundrucu samples from the south bank of the Amazons), as languages nearest in respect to their geographical localities—
East.—-The, Carib dialects, of which the nearest is Sir R. Schomburgk’s Guinau.
North.—The Maypure and Saliva families.
West.—The Zapara.
South.—The languages of the Province of Moxos.
To the Zapara the present vocabularies are theleastlike; and perhaps they aremostlike the Carib dialects. Now the Carib dialects have numerous affinities—some of them of a very remarkable kind. In the first place they have them with the Maypure and Saliva; next, they have them with the Guarani;thirdly, there is the statement (accompanied by a table which partially verifies it) in the ‘Mithridates’ that the Moxos and Maypure have several important elements in common.
The following Tables are formed from a very slight collation, viz. from one with short specimens of Guiana languages by Sir R. Schomburgk, and the equally scanty lists in D’Orbigny’s ‘Homme Américain.’ They are laid before the reader less on account of their own merits than for the sake of counteracting the common (but erroneous) belief that the South American tongues are wholly unconnected in respect to their vocabularies. The extent to which the commoner roots occur in different languages may be surmised from even so short a table as the subjoined.