Materials-Remarks

а corresponds to the normal or pure a, as it is in the Slav languages, and more or less like a in father (the length of the vowel not being taken into account).

е – is the open e, identical with the same vowel in the Slav languages and in German; like let in English (only observing that in English this sound is never final, but always followed by a consonant or vowel-glide).

о – is open, as in the Slav languages and in the German word Stock; like stock in English also (with the same remark as was previously made as to e).

i – corresponds to the high-front narrow i, like i in the Slav languages, in French, and the long German i (and English e), without taking into account the length of the German and the English vowel.

u – is a middle u, neither high nor low; more or less similar to the German short u, the Slav ordinary u, and the English short u (or w): good, foot, pull, bull, put.

In unaccented syllables, the vowels, o and on one hand, as well as e and i on the other, become nearly undistinguishable. For instance, o and e are apt to become narrowed, and on the other hand, u and i receive a more open pronunciation. Thus, certain authors write Ainо instead of Ainu, inao for inau, Ivаnki is also used as Ivаnḱе; с́ахḱе and с́ахḱi; śеrеmа and śirima.

This explains the sound y, which is occasionally heard, and is akin to the Russian y (ы): it may be defined to some extent as a non-palatal, 'hard', low i; or as a narrow e: somewhat akin to will, whisky, pronounced by a Scotchman.

It must be noted that in general the localization of the articulated vowel-sounds in Ainu is rather broad. This opinion has also been confirmed by Abbé Rousselot, who points out that a, e, o, in Ainu, may have the three different qualities of tone (timbre) found in the French vowels; i and u have only the two which are found in these. The diagram is as follows:

He is convinced that this difference of 'timbre' in vowels depends on their quantity. But it seems that the Ainus make no fixed distinction between short and long vowels; that is, they know nothing of quantity, properly so-called. We can only say that an accented syllable is longer, and may be simply termed long; but this length is in strict connection with the accent.

However, we do meet with fixed differences in quantity in certain words of which the sound would otherwise be the same; their only distinctive quality is the length of utterance. For instance: ē, 'to eat'; and ĕ, 'to come'; rū, 'the way', and rŭ, 'the ice of the river' or 'a flock of birds'.

As to the manner in which these differences have arisen, I am completely in the dark.

There are also (as in most languages) words that differ in meaning, though pronounced exactly alike, Ex: ki, 'grass', 'to do', 'a Blouse', and 'fat (subst.)'. ma, 'to swim', 'to fry', 'woman' (used in ccmps. instead of mat).

Lastly, there are quantitative differences which are used to lay (or to remove) psychological stress upon (or from) a word, thus increasing (or diminishing) its intensity of meaning. For instance : tеkоrо = very much ; tēkоrо = very, very much; оj = many; ōj - very many.

The Explosives are k, t, p; g, d, b. These two groups are not irrelated. In Ainu there is really only one group; if the sounds occur at the beginning of a word, their normal sound is k, t, p. In the middle of a word, the sound wavers between the former, the voiceless group, and the voiced group g. d, b. Strictly speaking, these are not identical with their Indo-European corresponding consonants. They are, I should say, neither "fortes" nor "lenes"; they are between. And then — which is yet more important — their conditions of combination. For instance, after m, these consonants readily acquire a certain sonorousness of tone, which probably does not last during the whole time of their articulation. The outcome of this was that in very many cases I was unable to determine the nature of the consonant, as I heard a sound that could not be identified either with the former group or with the latter.

At all events, amongst the Ainus of Saghalien, the normal and primary group is k, t, p, (voiceless), possibly less strongly articulated in certain connections. Their corresponding sounds (g, d, b,) more or less voiced, appear only as secondary variations. On the western shore of Saghalien, the latter group is more often to be met with than on the eastern shore.

There is a series of palatals, or rather palatalized consonants: ṕ, b́, ḱ, ǵ. (I use this mark (') above short letters and after long ones (according to their shapes), to denote the palatalization of the sounds). This palatalization wavers from zero upwards to the most distinct palatal sound. But I must say that Abbé Rousselot has come to the conclusion that k is not palatalized at all and that my Slav ear was at fault in this case. However, the k without (') might be read both by myself and many others like the hard k in the Polish kędy or the Russian кэта ('a kind of salmon'); this last word being often heard in Saghalien, and serving me as a term of comparison with those in which ḱ occurred. I therefore, without calling in question the learned Professor's opinion, have nevertheless continued to use ḱ in these cases, supposing there may be a difference in pronunciation between the Yeso dialect and that of Saghalien, which he had no opportunity to examine. At any rate, ḱ is never found but before e, and frequently very faintly palatalized then.

Nasals.

The nasals are ŋ, n, ń, m.

ŋ, which is very rarely met with, is (approximately) the English ng in hang, sing, long.

ń (marked in the Polish way) corresponds to the French and Italian gn (champagne, signor). The palatalization is weak, however, and at times imperceptible. It is found either at the end of words, or before e. Ex: niń, 'to dry up', etunńe, 'to dislike'.

Spirants.

The spirants are: s, ś, f, υ, h, x.

s is like the English s in soft.

ś has no equivalent in English. It is represented in most Ainu dictionaries by sh; Batchelor (Grammar, p. 77), gives it as equivalent to sh in ship. But this is certainly wrong It corresponds to a sound common in Slav languages, and represented in Polish by the letter ś; it is between s and sh, with a distinct palatalization, or an approach to the position in which i is articulated. Abbé Rousselot remarks that s is formed by the tip of the tongue held somewhat downwards, and its upper surface (dorsum) raised towards the palate; that sh (the French ch) is produced with the tip raised, the upper surface (dorsum) somewhat drawn in, and the lips thrust forwards; and that between those two positions there are an indefinite number of others wavering between s and sh, but never with the lips thrust out, as for sh. These positions correspond to the sound (or sounds) represented by various forms of the Slav ś. If we take ś and sh as two extremes, the Ainu ś lies between them, approaching now the Russian ś (sъ), now the German ch (in ich), now the Polish ś, or a palatalized š. And this explains Batchelor's remark that it is often difficult to know whether the sound is s or sh (Grammar, 78).

In the writer's opinion, we have two classes of words: one that always has s pure (mostly in the syllables sa and so), and the other in which there is a greater or less amount of wavering between the s pure, and the ś. Examples: iśam, there is not'; śumari, 'a fox'; the latter is often heard nearly as sumari. When I heard it thus, I wrote it down with s pure.

The origin of this ś is sometimes due to the influence of a foregoing i. Ex: sam, 'to marry'; iśam, 'to marry me'. Now and again the sound x becomes ś, as: Ipoxni (a proper noun), and lpośni; Inunuxnaj and Inunuśnaj (the name of a river).

s, so far as I am aware, is always voiceless. I noticed that the Ainus always pronounced Russian words in which z occurred, as if it had been s.

f occurs but seldom, and as a secondary sound, produced by the influence of the neighbouring vowel u. Examples: fura, 'smell', furaje to 'wash', kuf or kux or kući, 'a girdle', nuf or nu, 'a plain, a tundra'. Sometimes p, when weakened, becomes f, but always accompanied by u; ćup or ćuf, 'the sun or the moon'; tup or tuf, 'to change one's place'. - υ is always voiced; it occurs only at the beginning of words and between two vowels. Examples: υen, 'bad'; iυan, 'six'; ḱeυe, 'to drive away'. In some words υ takes the place of u: υare for uare, 'to cook'; haυe for hau, 'the voice'.

h has the sound of the English aspirate h. It is about the same as the German h, which is a glottal spirant, naturally voiceless, but which in combination, may become voiced.

x corresponds (more or less) to the German ch, or to the Spanish j and x. At the beginning of a word, x is rare, uttered more feebly, as in xunara, 'to seek'; xonive, 'an oar'; in the other cases, it is quite identical with the above-mentioned gutturals. Examples: xoni, 'the belly', ax, 'a fish-hook'.

Affricates.

These are: č, ć, ź.

The Ainus have no sound of c as ts, like the Polish c, or the German z: c is always somewhat palatal. There is also no pure č (ch in chalk). The sound always vacillates between a palatalized č, and something akin to the sound of the Polish ć (peculiar to that language).

There is no sharp line of demarcation between the palatalized t and the distinct affricate, like the Polish ć (which consists of ť + ś). It may therefore very well happen that from individual to individual, or from one dialect to another, there may be differences in pronunciation. This explains why the Russians, who as a rule have no affricate ć, have represented this sound amongst the Ainus by t (of course = ť), as Dobrotworski has done (Ainu-Russian Dict. Preface, p. 63; for inst. rame 'a house'). And again my friend Mr A. von Friken, Instructor in Agronomy, writes the word ćiśe with тисе, and ćoxća with тьохтя. for the same reason, in a small dictionary compiled by him.

There also exists a sonant variation of ć, the palatal ź, which is only met with in very few words, after a nasal consonant. Example: unźi, 'the fire'; tunźi, 'the interpreter'.

Liquids.

I have not met with l amongst the Ainus of Saghalien. Therefore they always prononce it like r in the Russian words in which it occurs. For instance, for молоко (milk) they would say мороко, хлеб (bread) they said xreb. But amongst the Ainus of Yeso, and in the village of Shiravoi more than in Piratori (the only two villages that I am better acquainted with) I have quite clearly heard l, pronounced by many persons instead of r. Neither Batcholor. nor any other writer of Ainu dictionaries (J. Dixon, The Aino Language. Chrysanthemum Magazine. 1883; or J. Summers, An Aino-English Vocabulary. Trans, of Asiatic Soc. of Japan 1886; and Furet, Vocabulaire Aino, Hakodate) have remarked the fact; but Abbé Rousselot's investigations confirm this observation of mine. He says: I exists only as a modification of r; the tongue, having its tip pressed against the palate, is able to vibrate along its edge.

R has certain peculiarities in Ainu. Dobrotworski, in his paper on the transcription of Ainu words (Preface to the Ainu Dictionary) says that r at the beginning of words has always a guttural (uvular) timbre, and besides is mostly heard as tr, which means that before the r there is an occlusive element. He represents this occlusive element by t, yet distinctly remarking that this t is articulated considerably farther up than the usual t. And he distinguishes three variations of this: 1) When t and are equally distinct, 2) When the 'velar' t is more distinct, 3) When the guttural r is more distinct. Whilst he gives such elaborate details of these variations, he adds that they all may occur in one and the same word. Professor Rousselot finds that r between vowels has its normal sound, but that at the beginning of words r is, as it were, semi-occlusive. By means of an artificial palate, he obtained the following variations: r, tr, kr, tl; and after n: — dr, dl, r, l, d. He says that the artificial palate showed the spot which characterizes an occlusive produced behind the teeth: which would explain the tr; and a contact of the velum in the place where k is produced: which explains the kr.

This contact, however, does not always take place; when it fails, only r or I is heard. He advised me notwithstanding, for the sake of simplicity in the graphic signs, to represent all this by r.

For my part I only noted down tr or dr when I heard these with absolute distinctness (which only occurs after n), and in such cases, I wrote a very small t (or d) above the line; for instance, pen^dram, 'the trunk (of the body)', ken^tram, 'famine'. I had therefore no difficulty in following the learned Professor's advice.

There is also to be noted the frequent interchange of t and r, especially amongst the Ainus of the North, according to my own observations. The naturalist Schmidt has also made the same remark1).

* * *

Certain final consonants are not completely articulated, and only very faintly heard. Ex: tarap, ćup. Words that have a final vowel often end in a quite indistinct manner. On account of this, a part of the preceding vowel is blurred in pronunciation; sometimes more, sometimes less. In such cases the vowel is breathed out like a weak x, which I have indicated by a small letter (x), or sometimes with the sound of an f, that I have designated in like manner (^f). It is probably in connection with this fact that the same word may terminate in sounds that have no relation to each other. As, v. g. ćup and ćux, 'the sun'; ćep and ćex, 'a fish'.

Accents.

The Ainu language has not the same place for the accent in all words, as for instance the last syllable in French (e mute apart), and the penultimate in Polish. In the following texts, we have, in order to simplify the printing, omitted to place an accent on any oxytone word (accented on the last syllable); for I have found that a very great many words are thus accented. Monosyllables on which stress is laid have been marked with an accent; so that, where there is no such mark, there is no stress on the word.

There are, as in English, primary (stronger) and secondary (weaker) accents (stresses); as, congratulation. In these texts, the primary accent is marked thus (´); the secondary, thus (`).

The accents are in general invariable for the same word. But there is a certain loftier style, used in poetry, prayer, public debates, receptions of guests, etc., which is called Sa koro ita (speech with intervals). These intervals or pauses, which occur after every two or three words and are represented by commas in our texts, render the style highly rhythmical, and often cause the tonic accent to be removed from the place that it has in common language. For instance, on line 6 of tale Nr. 1, the word oròpeká, is accented on the last syllable, whereas, in common speech, the accent is on the penultimate: oropéka. And again, 1. 119 of the same, we find the same word bearing different accents: noxrú, nóxru. And (lines 89 and 90) mintará — míntara.

There are, however a very few words which cannot change their accents without a change in meaning; v. g., átaj, 'a chair', atáj, 'payment,' án koro, 'my, his' and an koró, 'I have', 'he has'; śíri, 'the earth', śiri, 'a payment'; úma, 'a horse', umá, 'also'.
One pecularity of the Ainu accent (mostly in Saghalien) is that in continued speech the last word of the sentence — i. e., the word which precedes a pause — is accented on the last syllable, whilst the voice often rises, instead of falling.

We have marked with a special sign (~΄) a similar kind of pitch, by which the voice is raised and then falls suddenly, with an acute tonic accent. See 1. 79 in Nr. 1: okajãn΄te. This takes place when the speaker, having ended a sentence, suddenly adds te, by which the final verb is changed into a participle, and the former sentence joined to the next, with a complete change of construction.
A longer syllable will be marked (¯); a shorter one (˘).

The use of Hyphens.

It has been found impossible to nse hyphens with perfect consistency in these texts. We often cannot, in languages of such structure, distinguish with absolute certainty between suffixes and members of a compound word. Consequently we are unable to make a distinction between words that are homogeneous (whether for the time being or permanently) and mere group words. Ex: on one hand tu-o-hajne, 'to fear'; on the other tubiśkan, 'all'; mirokupo, 'a girl'; ankoro—kanne, 'having', etc. At times it has been necessary to follow the etymological divisions; at others, those of actual speech. For instance, in ćisé-koro, 'house-possessing' the hyphen serves to denote the fact of the two last syllables having no independent accent. In reading, we have to distinguish, v. g., between or-o 'to put inside', and oro 'from'.
I have written the often-recurring word an koro, as an-koro, when it meant my, his, her, and ankoro when it signified I have. I have dealt with an in the same way, when it occurred in other words.

Open	Medium	Narrow
à	a	á
è	е	é
ò	о	ó
	i	í
	y	ý