In speech, as in everything else, the forest Indian is confined within the narrow limits of his immediate surroundings. Unlike the nomadic Indian of the plains, he passes his entire existence in an area little larger than an English parish. He has almost no commercial dealings with his neighbours. The only fresh blood that penetrates his tribe is brought in by the immature children taken prisoners in war. Like the landscape his imagination owns no perspective, no horizon. In the Amazonian bush an Indian may live and die without ever having gazed upon a terrestrial object at the distance of a mile. His mode of life, a community within a single house, under a single roof, makes of household words a dialect, and with the passing of a generation makes that dialect a language.

In a society where each tribe is complete in itself and at deadly enmity with all its neighbours, and where writing is unknown, language must naturally undergo very rapid, very definite change. Moreover Indians will not voluntarily speak the language of other Indians. Thus the Amazonian natives use no common tongue, and there is little in the vocabularies so far collected to explain either the origin or the relationship of the existing dialects. Tribes divided by the breadth of a narrow river speak languages that are mutually unintelligible. On the other hand, tribes distant some hundreds of miles from each other possess a language with a common root, which is fundamentally different from those in use among all the intervening peoples.

So far as I could classify them, the language-groups of this district fall under thirteen headings. By group I comprehend all tribes speaking a language with common roots, though the dialects may vary considerably. These groups, and the approximate number of Indians in each, are as follows:

According to Koch-GrÜnberg all the tribes on the Tikie speak the Tukano language, and as a result of segmentation the Airi and Tihio speak the Dessana language.

Occasionally tribes, though speaking an entirely diverse tongue, and members of a distinctly different language-group, have some comprehension of the tongue spoken by a neighbouring language-group. For instance, the Muenane can understand Witoto, but they have no knowledge of Boro, probably because they come more in contact with the former people. The Menimehe know some words of Tupi, or lingoa-geral, which is extraordinary, even though their acquaintance with it is very slight.

The tribal names in ordinary use are, as has been said, bestowed by neighbouring tribes, and are merely nicknames. It follows that the name by which a tribe becomes known to a traveller is the name in use among the tribes in the districts through which he passes, so that a visitor from the north probably knows of a tribe by a different extra-tribal name from that known to a new-comer from the south. The difficulties of identification caused by this have already been commented on in an earlier chapter, it is only necessary to refer to them here in so far as the same difficulties beset any attempt to learn the local dialects.

Of the thirteen languages tabulated above, one of the most difficult, and the most guttural, is the tongue spoken by the Resigero group of tribes. Nonuya, also guttural, is perhaps equally difficult, whilst Andoke is possibly the worst, as it is almost ventral. Okaina, though presenting many difficulties, is easier to acquire than the first-named three, and may be characterised as nasal, while Boro and Witoto are neither nasal nor ventral nor impossibly guttural. Muenane is somewhat akin to Boro, but is richer in words. Menimehe approximates more to the speech of the Uaupes River Indians, and it again is nasal.

The endeavour to reproduce the guttural expressions of the Indian in Roman letters is rendered the more complex by the uncertainty of his utterance and the aural variations of his European interpreters. The same word phonetically transcribed by an Englishman, a German, a Frenchman, and a Spaniard bears little or no resemblance to a common inspiration. Each European observer conveys to his written word the error of his national idiosyncrasy of impression and pronunciation.

The difficulty of a phonetic rendition of a foreign language into English has long been apparent, and is one shared—though in a lesser degree—by all Continental linguists. To meet this difficulty the Germans have devised a system almost Chinese in its intricacy, while the French seek to reproduce such simple sounds as that of our English “W” by combinations of diphthongs. Many of these elaborate phoneticisms have been adopted by English writers without consideration of the lingual limitations of their inventors, or of the confusion induced in the mind of the student.

To simplify transliteration, though at the sacrifice of the finer distinctions of the language, the orthographic system of the Royal Geographical Society has been used in this work,[402] and the explanation of the system given in the appendix with the Witoto and Boro vocabularies is taken from the rules laid down by that Society and adopted by the Royal Anthropological Institute.[403] This system ordains that an approximation to the sound should be aimed at only, as any system which attempted to represent the more delicate inflexions of sound and of accent would be so complicated that it would merely defeat itself.

I attempted to make a vocabulary of Andoke words, but the language is, as I have noted, so guttural, not to say ventral, that it renders all attempts impossible without some medium to work upon at the start, such as I had with Boro and Witoto. In these two cases Brown’s knowledge of the latter, and even his very slight acquaintance with the first, were of great use to me as a basis upon which to work.

As an example of the difficulty to be faced without some common medium, I have asked a native, “What is this?” and touched my head or a stick, but could find no clue to whether his answer referred to the thing touched or my action in touching it. Only a long and tedious study can overcome conundrums of this description, and when to these is added the impossibility of conveying accurately by written signs the sound as uttered, the attempt proved beyond my powers and resources.

Mention has been made by one writer of the “‘cluck’ of satisfaction—common to all the tribes of the Provincia Oriental.”[404] I consider the sound emitted by the Issa-Japura peoples as a sign of assent or pleasure is more ventral than that described by Simson. It is approximately Hurrr! like a grunted sigh of satisfaction. The exclamation of surprise amongst all these peoples is very similar and may be written Huh! This sound, lengthened considerably, is the Witoto affirmative Huhhh. Huh! huh! huhh! as affirmatives are very freely used in conversation. The more an Indian agrees with the speaker the more ventral do his ejaculations become. The negative will not be used except in direct answer to a question, for it is contrary to Indian custom and etiquette to interrupt or contradict. The absence of the affirmative Huh! is practically a contradiction, on the ground of doing nothing being itself negative. A similar idea is seen in the tobacco palaver, where the dissentient signifies his disapproval by abstaining from licking tobacco. Should an Indian, however, wish to give an affirmative answer to a negative question, he will then make use of the negative No, for to answer Yes in Indian parlance would be to confirm the negative.

This brings us to the question of construction, and it is at once apparent that in Witoto, for example, the construction of a sentence tallies more with the construction of the deaf and dumb mute’s gesture language than with anything else, that is to say it is the very antithesis of the Chinese, or of our own. It may be said of the Indian, as Tylor wrote of the deaf and dumb mute, that he “strings together … the various ideas he wishes to connect, in what appears to be the natural order in which they follow one another in his mind.”[405] For instance the Witoto say, Benomo honne, literally “here put it”; benomo ekkono, “here open it you”; eijo rie dotoenyino, “much fruit put in it not do you”—“do not put much fruit into it.”

It will be noted that the personal pronoun here has become the suffix of the verb. This is the general rule, as in dinitikwe, “I shall carry it”; a chimitekwe, “I am going to see”; ona dueruetckwe, “I want you.” But this rule is not invariable, as we find kwe mona, “I am unable”; ke hanyete, “I do not understand,” with the pronoun kwe or ke placed, as we should put the “I” before the verb; nor is the variation caused by the negative, as “I do not want you” is ona dueruenetckwe. In this instance the position of the personal pronoun kwe is probably determined by the objective ona, which structurally must precede, otherwise the meaning of the phrase would be inverted and become “You do not want me.” A pronoun is also used as a prefix to a noun to denote possession, as tano, “cassava,” ometano, “your cassava.” According to Koch-GrÜnberg the suffix make indicates some other place, or thing; it occurs in baimake, “other”; naimake “them”; but I am not aware that it acts as a definitely differentiating suffix in these or any other case.

In Witoto nyete as a suffix negatives what proceeds it, the literal meaning of the word inyete—a compound of ite = are—being the equivalent of the French il n’y en a pas.[406] As examples of its use we get figora, “good,” figonyete, “bad”; huchite, “twisted,” huchinyete, “straight,” that is to say “not twisted.” The Boro negative is ne, as for instance in imine, “good,” nemine, “bad,” i.e. “not good.”

Repetition of a word literally doubles its meaning, as in the Witoto nana, “all,” and the Boro paa-paa, “low-low,” that is to say, “lower”; kame-kame, “high-high,” i.e. “higher.”

I have said that the principle of construction in both Boro and Witoto is that of the mute’s gesture language, but gesture language actually is almost unknown, non-existent, among all these tribes. The hand is pointed to show direction, or to identify a person or object. The Indian beckons with one hand, but its movement is downward, not upward as with us. There is also a recognised sign to express desire for sexual intercourse. This is a mere jest, a ribald suggestion, as with boys of a certain age among our lower classes. The right elbow is grasped with the left hand, the elbow being so flexed as to allow the hand to point upwards. It is, in fact, the letter Z of the dumb alphabet.

Fingers and toes are used for reckoning, and are the more needed in that the Indians’ knowledge of numbers is of the slightest. But few can reckon beyond five, though I once found a senior wrangler who counted seventeen, by the aid of all his fingers, all the toes of one foot, and two of the other. The remaining three toes he covered over, to show that they were not required for the total sum. If an Indian wished to enumerate anything over ten he would place both hands to his head and say, “Like the hairs of my head.”[407] In Boro I could only learn of four numerals, tiamie, “one-half”; tsanere, or tsape, “one”; mieke, “two”; sause, “five.” These in combination give tsape-mieke, “three”; mieke-mieke, “four.” The Witoto numerals are dahe, “one”; mena, “two”; dahe-amene—equivalent to the Boro tsape-mieke,—“three”; menahere, “four”; dapekwiro, “five”—that is one hand; nagapekwiro, “six.”

It makes absolutely no difference to the value whether you say tsape-mieke or mieke-tsape; dahe-amene or mena-dahe.

For measures these tribes have nothing more definite than a handful, a foot- or finger-length, and of weights they possess no knowledge whatever, nor, so far as I am aware from their customs or their language, is there any consciousness of more possible or desirably-accurate definition.

To express a length of time other than the merely immediate past, present, and future, the Indian makes use of what conveys to him an indefinable idea, “As long as the hairs of my head.” This is similar to his notion of expressing any large number. He reckons time by the moon to the extent of saying, “When the moon is small,” or, pointing to it, “As it is now,” but I never heard anything like “so many moons,” or an equivalent value in a word. In fact, time to the Boro, so far as I am aware, is distinguished by only pekare, “to-morrow,” aiupe, “yesterday.” The Witoto will speak of beiruito, “to-day”; wiremoni, “to-morrow”; dawire, “the day after to-morrow,” or nawire, “yesterday”; beinawire, “the day before yesterday,” or beinawife, “the night before last.”

There is, as I have already mentioned, no writing, not even the most primitive picture-writing. The Indian makes use of no signs as aids to memory; and the only recognised symbol that I met with—other than such symbolic practices as the presentation of wood and thatch by the bridegroom to his parents-in-law—was the tobacco folded in a strip of palm leaf that is the regular invitation card of North-Western Amazonia when festivities are toward. Neither individuals nor families have any recognised name-marks—such as a peculiar notch or number of notches—to distinguish personal property. It must be remembered that in the small private habitations in the bush a man and his wife and children are more or less isolated, and that in the great tribal house the family community have most of their possessions in common. It is difficult with so communal a people to know what may be looked upon as general property, and what as individual, with the exception of personal ornaments. Indians recognise their property only by differential qualities, certain ornamentation, ways of binding or lashing, patterns in basketry, colouring—and division of colours—on pottery; and these differences are known and recognised by others, as well as by the actual owners.

Each tribe has its peculiar call or signal, which I believe is altered occasionally as a precautionary measure. This may be a whistle, or the imitation of the cry of bird or beast. Then there is the so-called drum-language used in signalling, and already noted in a previous chapter, which I certainly believe to be some sort of code. Brown’s assertion that the sound of the word is made with the drum, and the Indians’ description of making the words is, I take it, merely the untaught intelligence striving to explain how an onomatopoeic language—such as Boro and Witoto to some extent certainly are—can be further conventionalised to a scope even more circumscribed than the ordinary monotone of the Indian’s speaking voice.

Not only is the Indian voice monotonous, but the conversation is rendered yet duller by the invariable repetition of the last words of a sentence. This is particularly the case with the Tuyuka, where conversation has a definitely ceremonial form. For instance, if a man leaves a party to bathe, he says, “I go to take a bath,” and the company present reply in chorus, “You go to take a bath.” On his return the formula runs, “I have taken a bath,” and the confirmative echo follows, “Yes, you have taken a bath.” This endless repetition, as was noticed with regard to songs, is characteristic of all Indians.

In quality their voices are strident and rasping, and are always raised in conversation and grow higher with increased excitement. No Indian speaks confidentially, he shouts; and unless something very sacred and secret is under discussion the conversation in an Indian house can be heard a mile away. In the forest the mass of vegetation above appears to act as a sounding-board, and so to lengthen the distance that sound is carried, not, as one might think, to stifle it. But independent of this the Indians possess extraordinary power of throwing the voice, a sort of ventral whisper; and all, to some extent, are ventriloquists. Even semi-civilised Indians of Brazil, who have lost much of the cunning of their brethren, the “Wild Indians” of the forest, have this power.

The Indian is as fond of speaking and singing in a high-pitched voice interspersed with ventral grunts as a Chinese coolie, and this predilection, as regards the falsetto voice, is greatest on the part of the women, whose voices are always higher than the men’s.

When an Indian talks he sits down, no conversation is ever carried on when the speakers are standing unless it be a serious difference of opinion is under discussion; nor, when he speaks, does the Indian look at the person addressed, any more than the latter watches the speaker. Both look at some outside objects. This is the attitude also of the Indian when addressing more than one listener, so that he appears to be talking to some one not visibly present.