Burushaski An Extraordinary Language in the Karakoram Mountains Dick Grune, [email protected].

nl

Introduction Bur shaski (stress on the second syllable) is spoken by some fifty thousand people in northern Pakistan, the Hunzakuts, who live in the valleys of the Hunza and Yasin rivers (north of Gilgit) where these cut through the Karakoram Range [lit.ref. 5]. The language is not obviously related to any of the surrounding languages: the Indic languages of Pakistan or the Tibetan languages of China and northern Kashmir. Although Burushaski has been compared to almost any language on earth, no fully convincing relationships have yet been established. Modern taxonomic methods are, however, beginning to yield results. Ruhlen (1989) [lit.ref. 7] still classified Burushaski as a language isolate: its genetic affiliation remains a complete mystery (p. 126), but Ruhlen (1992) [lit.ref. 7] reports on a possible classification of Burushaski as a separate branch of a newly proposed Den -Caucasian superstock. More recently, Blazek and Bengtson (1995) [lit.ref. 8] list tens of etymologies relating Burushaski to the Yeniseian languages, spoken by a hundred people along the Yenisei river in Siberia. Where appropriate, we have included these etymolgies in this survey. These etymologies are not (yet) supported by full reconstructions; to emphasize their tentative nature, we show them between square brackets. The proposed Dene-Caucasian superstock comprises Basque, the North-Caucasian languages, Burushaski, the Yeniseian languages and the Sino-Tibetan languages in Eurasia, and the Na-Den (Atabaskan) languages in North America. Blazek and Bengtson [lit.ref. 8] list 219 etymologies, some very convincing, others less so, to establish its existence. The group is estimated to have a time depth of more than 10000 years, and is thought to represent some (or perhaps all) of the "old languages" of Eurasia. Burushaski exists in two dialectal variants, Hunza and Y sin (= Werchikw r), spoken in their respective river valleys; both rivers run north-south from the Karakoram Range the Hunza about a hundred kilometers east of the Yasin and flow into the Gilgit river. The differences between them seem smaller than those between the London and Liverpool variants of English (but larger than those between British and American English). Speakers of Hunza and Yasin Burushaski have little difficulty understanding each other. For an impression of the differences, see the chapter Numerals: Hunza Burushaski and Yasin Burushaski compared. Lorimer [lit.ref. 4] contains extensive comparisons. This survey is based on Yasin Burushaski, mainly because for this variant the most modern material is available [lit.ref. 1,2,3]. From the sometimes hesitant nature of the reports and from the differences between them it seems doubtful that together they describe the grammar of Burushaski completely, and it is obvious that the dictionaries are far from complete. The definitive work on Burushaski has not yet been written and much research remains to be done. Although many adult speakers of Burushaski also speak Khowar, Shina or Urdu, Burushaski is the normal means of communication in both the Hunza and the Yasin valleys; it was and still is the carrier of a vigorous story-telling culture. Burushaski is not an endangered language. Until recently, Burushaski was not a written language and original historical material is very scarce. Some texts have been written down by native speakers at the request of researchers; these texts are in the Urdu variant of the Arabic/Persian script. Lorimer [lit.ref. 4], part 2, contains many texts in the International Phonetic Alphabet. Berger [lit.ref. 1] and Tiffou [lit.ref. 2] supply more than a hundred pages of text in romanization; we will use almost the same romanization in this survey. With the advent of the Karakoram Highway along the Hunza valley, the Hunzakut society has opened up, and literature is now being written in Burushaski [lit.ref. 5]. Due to external influences, more than half the present-day Burushaski vocabulary is of Urdu, Khowar and Shina origin (Khowar and Shina are two Northern Indic (Dardic) languages, closely related to Kashmiri and, somewhat further away, to Hindi/Urdu). It is the rest of its vocabulary and its structure that make Burushaski a language isolate. There are perhaps half a dozen loans from Turkish and one or two from the Tibetan language Balti, spoken on the other side of the Karakoram range.

-2-

General Structure and Phonetic Features For all its romantic and exotic associations, Burushaski is not much weirder than Latin, Turkish or Finnish; of these three it is most reminiscent of Turkish in its structure. It has two or three cases for the nouns (see below) and a small number of locative suffixes; it has essentially one conjugation for the verb, plus a number of composite conjugations; and its sentence structure is similar to that of Turkish but much simpler. Its most remarkable features are that it has four genders for the nouns and that the indications of the object of the verb are the same as those for possession on the noun: I hit him is expressed roughly as I do his hitting, as in many Amerind languages. p - as in French p re (i.e., voiceless) ph - as in English pair (i.e., aspirated) b - as in English bear (i.e., voiced) t - as in French tant th - as in English tan d - as in English Dan t - retroflex t (i.e., with the tongue curled back) th - retroflex th d - retroflex d k - as in French quand kh - as in English cant g - as in English gone q x - the Arabic qof - as in Scottish loch - a voiced version of x Figure 1 Plosives in sets of three
 

Phonetic features The consonant system of Burushaski is rather similar to that of Hindi/Urdu, except that the voiced aspirates (bh, dh, dh, gh, jh) and the retroflex n (n) are missing. Most consonants in Burushaski come in groups of three: voiceless, aspirated and voiced. The plosives are shown in Figure 1, the sibilants (s-like sounds) in Figure 2. This arrangement in groups of three is significant in the language, since there is a phonetic change, called hardening, which turns the third consonant of each group into the first (and which does not affect the others); that is, b turns into p, into q, j into , etc. For one situation in which hardening occurs, see under Negation c below. The other consonants are: h, m, n, ng, w, y - as in English l - as in German, but often almost voiceless, especially before t r - rolled, as in Spanish In some words that start with h- the h hardens to -k-. It is possible that the frequent combination -ltshould be considered as a single consonant. Burushaski has the usual five vowels a, e, i, o, and u. Stressed they sound like their Italian counterparts, unstressed they sound like the English vowels in but, bell, hit, hot, and put, as they would in Hindi. The great majority of Burushaski vowels is short, but stressed vowels can be long; a long vowel is marked by a circumflex by Berger and more scientifically but less conveniently by a macron by Tiffou. We use the circumflex: ne = he, n = him-of = his. Occasionally an unstressed vowel is found to be long, or at least to have the (Italian) sound of a stressed vowel. In unstressed position, the difference between e and i tends to disappear, as it does between o and u. Berger writes g smu = of a woman where Tiffou writes g smo. If the first syllable is unstressed, the vowel in it tends to disappear at all: Berger writes cer - = to go, Tiffou writes cr  

c - English ts s - English s z - English z

c - retroflex c s - retroflex s j - retroflex j Figure 2 Sibilants in sets of three

c - like English ch s - like English sh j - like English j

-3

(pronounce tsre-, not kre-!). Also, the word Burushaski is often pronounced Br shaski. Stress is fairly prominent, more so than in Italian but less than in English; its normal position is on the second syllable (as in Bur shaski), but all other positions are possible. Stress is indicated by an acute mark () on the vowel, except in words of one syllable which always carry stress; there are no small unstressed words such as the English an in Burushaski. Since the nature of a Burushaski vowel is determined by four factors, position (a, e, i, o, u), colour (what was called Italian versus English above), stress and length, a complicated picture emerges that is interpreted somewhat differently by the different authors. It would seem that the original Burushaski vocabulary has no long vowels, and that long vowels result only from two equal short vowels running together. Indeed n = his can be interpreted as n -e = he-of, analogous to h r-e = man-of = of a man from hir = man. Many other examples of this phenomenon can be found in this and other texts. In addition to these composite long vowels, foreign loans and onomatopoeic words can also contain long vowels; an example is p t- = to type (lit. to c do "chahp"). The Burushaski language section of the Lonely Planet Walking Guide "Trekking in the Karakoram and Hindukush" [lit.ref. 6] marks all stressed vowels as long. Many nouns and verbs require a prefix and/or a suffix and may then require the stress to be on either of them. These nouns and verbs are shown in the dictionary with the acute mark on the hyphen before or after the word: an entry skil = face means that the word skil = face requires a (possessive) prefix which will carry the stress; e.g. skil = my face. Concerning the Burushaski Noun Nouns are classified into four categories, similar to the genders in German or Latin; the criteria for the classification are somewhat different, though. The classification is based on three distinctions. The first and most important distinction is between countable and uncountable. Examples of countable things are hir = man, huk = dog, balt = apple, etc; examples of uncountable things are cel = water, da m = flour, ji = life/soul, etc. The group of countable things is called hx, that of uncountable things y, for reasons that will become clear in the next paragraph. The second distinction is that countable things are divided in human and non-human; the human class is called h and the non-human countable class x (all humans are countable, of course). Third, the human class is split into masculine and feminine: hm and hf. In total this gives us the following classification: Class hm hf x y Description human males human females animals and countable objects materials and abstracta Singular n m s t Plural w w c k
     

The class of a noun shows up in many conjugations and declensions, often through the presence of a specific consonant associated with the class; for convenience these consonants are given in the columns Singular and Plural. An example of this phenomenon is:
  

khen khom gus gut

this (man here) this (woman here) this (animal or thing here) this (material here)
   

in which the kh-/gu- part expresses nearness and the rest (-n , -m , -s , -t ) expresses the gender; a similar set exists for the plural. In general, the above classification is more in accordance with reality than that of the Latin nouns in masculine, feminine, and neuter. All nouns referring to human males are hm, all nouns referring to human females are hf, all animals are x and almost all materials and abstracta are y. For concrete non-animate things, however, the situation is less clear. First there are a number of words that can be both x and y, with the expected difference in meaning: bay = salt is y, and bay = lump of rock salt is x. This is comparable to Dutch de doek = the cloth, the rag versus het doek = the cloth, the material. Something similar happens with fruits: balt = apple is x (as expected), but the same word balt = apple tree is y, as if apple trees were a continuous material from which apples are the lump-like manifestation. Another glimpse at the underlying difference between x and y is given by - l (x) = intestines of living animals and - l (y) = intestines of dead animals. [Yeniseian has t l = intestine.] Some gender assignations are not so easily explained: ha = house, e.g., is y, although it is countable and not a material, nor abstract; ran = cream is x, although it is a material and not countable; and n us = stinginess is x, although is an abstract. Especially, small plants are divided 50-50 over x and y. Another, somewhat amusing fact is that female spirits and fairies are hf, but male jinns and monsters are x rather than hm; Xud = God is hm, though.
      

-4-

Noun forms Nouns in Burushaski are used in noun forms. A noun form tells something about the function of the noun in the sentence; it often corresponds to a noun phrase in English: in my hand, for the benefit of the town. A noun form normally has only one stressed syllable; some noun forms are so long as to have two stressed syllables and are then considered two separate words. The noun form consists of two parts: the noun part and the endings part. The noun part contains in order: possibly the owner (i.e., my, your, his, etc.), followed by the stem of the noun, followed by a number indicator, if applicable. The ending part is either a case ending (absolutive, oblique or genitive), or a location ending, optionally followed by a direction ending, or a pseudo-case ending. The main difference between the case endings and the pseudo-case endings is that at least one of the first group occurs in every sentence. An example of a noun form is ar na = into my hand, which consists of c a-r n-i-a = my-hand-in-towards. An example of a two-word noun form is: hare gand i = for the c s c benefit of the town, which consists of har-e gand -i = town-of benefit-in. s c Almost all endings are preceded by the hf indicator -mu- in hf words. We shall now look at each of the above components, starting with the number indicator. Number Nouns can be singular in number, indefinite, plural or grouped. The simple form of the word indicates the singular: ha r = the horse(x). The indefinite is formed by suffixing -en to h nouns and -an to xy nouns (compare the h and xy forms of the word for one under Numerals): ha ran = horse-some(x) = some horse. The meaning of the indefinite is some not yet mentioned X and it is used to introduce new topics: Hen h ren bam, ... = one man-some was, ... = Once upon a time there was a man, ... . The plural is formed in quite a diverse number of ways and is treated in the next paragraph; its use is equivalent to that in English: ha ra = the horses(x). The group plural (also called double plural) is formed by adding -ek to the singular or plural of the word: ha rek = groups of horses(x). This suggests that the real meaning of a plural like ha ra is group of horses rather than just horses. Another example is s sen = some person, ses = a group of people, and s sek = groups of people. Words of the y class can also have plurals, even if they denote materials: cel = water, celm ng = gushes of water, celm ngek = several gushes of water. The plural Very broadly speaking, the regular plural ending for hx words is -u and that for y words is -ing; hx words of one syllable get - nc. There are many particulars, however, and some two hundred nouns do not fall under these rules. Since the plural is difficult to predict and since the plural ending almost always tells you if a word is class x or class y (hm and hf are never a problem), the dictionary gives the plural of each word, thus supplying two pieces of information for the price of one. [Yeniseian nouns have a plural in -ng and a collective plural in -n.] The way the -u is connected to the noun for hx words depends on the final letter of the word. Words ending in a vowel get -mu: p qu = loaf, p qumu = loaves; those in -n have the -n replaced by -yu: dum n = s enemy, dum yu = enemies; those in -s have the -s replaced by -u: gil s = drinking glass (from s s English), gil u = drinking glasses; and the rest gets -iu: d st = friend, d stiu = friends. s s s Words of the y class get -ing, as explained above: asq r = flower, asq ring = flowers, except those ending in -i, which get -ng, unsurprisingly: pi = pliers, ping = pairs of pliers. c c In both classes there are irregular plurals by the dozens; Figure 3 shows some examples which will give an impression of the variety (in addition to the ha ra = horses shown above). We see that the plural of das n = girl is a word totally unrelated to das n (although it is clearly related to gus = woman). This phenomenon, which is called suppletion, is rare in Burushaski, as it is in English (I go versus I went is an example). Hunza Burushaski has no suppletion is this case: das n = girl, das wanc = girls. [Burushaski gus = woman may be related to Yeniseian qos = witch.]
                                 

-5-

Possession, Location, Direction The possessor of a noun is expressed by putting a form of the personal pronoun before the noun, as it is in English: my foot; but there the similarity ends. Figure 4 shows the various forms of the possessive pronoun prefixes in Burushaski. The unstressed i- is almost universally left out (indicated by the square brackets in Figure 4); the unstressed u- is also often omitted.


There are two types of nouns in Burushaski with respect to possession: those that can occur without a possessor and those that cannot; the latter are called inalienable. Most words, and certainly all foreign words can occur without a possessor, but most family relations and body parts require a possessor upon each and every use, as in many American Indian languages (for instance Navaho and Hopi, but also in many, many others). Although the dictionary says that the word for hand is -r n, the hyphen before it shows that this word always requires a possessor and that there is in reality no word for just hand. A speaker of Burushaski, asked What is the word for hand, will answer ar n = my-hand or gur n = your-hand, depending on where you point. Words that do not require a possessor can be given a possessor by putting the genitive of the corresponding pronoun (see Figure 5) in front of the word, much like the English my, your, etc.:
 

Most of the inalienable words get Type I prefixes; a few get Type II and only two get Type III. The Type I words come in two varieties, those that have the stress on the word itself (e.g. -r n (y) = hand) and those that have the stress on the prefix (e.g. skil (y) = face). So we get  

Note that we have ren = his hand, rather than ir n, since the unstressed i- is normally omitted. So, after all there is a word ren from -r n = hand, but it means his hand rather than hand. Often the genitive of the pronoun is put before these forms, for good measure: j ar n = of-me my-hand, n ren = of-him his-hand.
   

ar n gur n ren mur n etc.

my hand your hand his hand her hand

skil g skil skil m skil

my face your face his face her face

j g n m etc.
  

ha ha ha ha

r r r r

 

Figure 3 Examples of plurals

Possessive/Experiencer Type II g m m m          

of-me horse = my horse of-you horse = your horse of-him horse = his horse of-her horse = her horse

 

Figure 4 The personal pronoun prefixes

we you they (hmf) they (x) they (y)

mimauu[i-]

m m -

person I you he (hm) she (hf) it (x) it (y)

Type I agu[i-] mu[i-] [i-]

Type III g m -

Class hm hf hf h x y y

Singular hir - man gus - woman das n - girl gam ic - neighbour hal - fox ha - house cel - water

Plural hur (also hur kia) - men gu nga - women s gu ngia - girls s gam iu (, not c!) - neighbours c c halj - foxes hak ang - houses c celm ng - gushes of water

-6-

A very probable theory says that Type II words were originally Type I words that started with a short -a-, and that the -a- has become assimilated to the Type I prefixes to form the Type II prefixes. Basing oura selves on this theory, we mark the hyphen before the Type II words with a small a: -cu = sibling of the same sex (brother/sister). Type II prefixes always carry the stress; so with this word we get: cu g cu cu m cu etc.
    

my brother/sister your brother/sister his brother her sister

(Berger [lit.ref. 1] uses a hyphen with a dot on top: -cu which is difficult to distinguish from , and Lorimer [lit.ref. 4] writes the hypothetical a (aco for Hunza Burushaski) but the word does not really occur in this form). - A A The two nouns that get Type III endings are -skir = father-in-law and -skus = mother-in-law, in which the Type III has been indicated by a small capital A over the hyphen. This yields forms like skus = my-mother-in-law. One can suppose that these words derive from hir = man and gus = woman, resp., through an infix -s- which produces hardening of the following consonant. What the meaning of this infix would be is unknown. To emphasize ownership, ya = own can be used: ya ha = my own house. The word ya is also used in a number of idiomatic expressions, for example: ya bi - it is mine my-own it-is
 

Case

Case endings There are three cases: absolutive, oblique and genitive, but the forms of the oblique and the genitive are almost always the same, so the impression may easily arise that there are only two cases. The absolutive has no ending, and the oblique and the genitive are both marked by an ending -e. One difference between them is that the genitive -e is preceded by the hf indicator -mu- for hf words in the singular, and the oblique is not. Another is that the genitive of you is g and the oblique is un (see Figure 5). Also, older material and some frozen expressions suggest that the genitive ending was -o in the past. All this shows that the oblique and the genitive are indeed different cases. [Yeniseian has aj = I for Burushaski ja, and u, ku = you for Burushaski un, gu. This is one of the most convincing correspondences.] The -e disappears after unstressed final vowels, but lengthens stressed vowels; examples are shown in Figure 6. Many additional examples that follow the above rules can be found among the pronouns shown in Figure 5. The forms separated by / are alternative forms; the left ones are the more usual ones. When case endings are added to a compound consisting of more than one noun, only the last noun gets the endings, as can be seen in the last example in Figure 6. The third person pronouns double as articles, both in the singular: hir = a man, ne hir = the man, and in the plural: ha ra = horses, ce ha ra = the horses. In the oblique, only the noun gets the ending -e: ne
  

Figure 5 The personal pronouns

we you they (hmf) they (x) they (y)

mi ma we/u ce ke

m m w /u c k

m m w /u c k

person I you he (hm) she (hf) it (x) it (y)

abs. ja un ne mo se te

obl. j un n m s t

gen. j g n m s t

-7-

hire = the man (oblique). The absolutive is used for the subject of the intransitive verb and for the object of the transitive verb; the oblique is used for the subject of the transitive verb only. In other words, the absolutive indicates the experiencer and the oblique the actor of an event. This is profoundly different from the nominativeaccusative scheme Europeans (except the Basques!) are used to, which is why different terminology is used. Examples are: - the man (absolutive) fell ne hir w li ne h re - the man (oblique) mo gus muy ci saw the woman (absolutive) mo g se - the woman (oblique) ne hir y cu saw the man (absolutive) Now it would be tempting to say that Burushaski just has subject and object reversed, and that the verb -y cwhich the dictionary says means to see actually means to be seen, but this is incorrect: the verbs ending is determined in each case by the real subject of the sentence! In ne hir w li = the man (absolutive) fell, the final -i indicates an hm subject and derives from the absolute form hir = man (absolute), but the final -i in ne h re mo gus muy ci = the man (oblique) saw the woman (absolutive) indicates again an hm subject but derives from the oblique form ne h re = the man (oblique). And in mo g se ne hir y cu = the woman (oblique) saw the man (absolutive), the final -u on y cu = him-saw-she is in accordance with the hf nature of mo g se = the woman (oblique). Location and direction endings There are three location endings: -i c - in -ce - at, near, attached/pressed to the outside of -yate - on, with (= by using) and two direction endings: -a (-a after vowels) - to, towards -um - from Each of these can be used as is, or be combined with one of the other group, where the location endings come first and lose their final vowel. These endings are, like all endings except the oblique e, preceded by the hf indicator -mu- for hf words in the singular. The location ending -ce lengthens a preceding stressed vowel, -i c and -yate do not. Figure 7 shows some examples. The ending -ce has no direct English equivalent and often requires a special translation, as in:
             

 

the dog and the cat

huk ka bu s

huk ka b e s

huk ka b e s

Figure 6 Examples of case endings

my brother my sister

cu cu

cu cu

cu cumu

man speaking woman speaking -e to last noun only

a horse horses a cow

ha r ha ra bi

ha re ha ra bi

ha re ha ra bi

 

 

Noun a man a woman some man some woman women

Absolutive hir gus h ren g sen gu nga s

Oblique h re g se h rene g sene gu nga s

Genitive h re g smu h rene g senmu gu nga s

Comment

-e disappears indefinite number -mu with indef. no -mu with plural -e disappears stressed - lengthened

-8-

g stingce - around your waist your-waist-(attached-to)

Pseudo-case endings and composite noun forms There are two or three other less frequent case endings, of which we will only mention -ule = at (at a certain time): te g ncule = that day-at = at that day. And then there are dozens of compound endings, corresponding to English structures like on account of, c according to, etc. We have already seen hare gand i = town-of benefit-in = for the benefit of the s town; another is s nde it = river-of remote-side = at the other side of the river. Concerning Adjectives and Adverbs in Burushaski Adjectives behave to a large extent as nouns; there is an adjective u = good, but it is difficult to tell whether s it means good or a good person or thing. It may be used attributively (directly with the noun), as in se u s ha r = this good horse; but when used predicatively (in a statement) it is used as a noun: se ha r u n s bi = this horse a-good-one is = this horse is good. Many Burushaski adjectives end in -um, which may be related to the ending -um = from, which we have seen above. The nature of this relation may be shown by examples like harum ses = town-from people, s which can either be interpreted as people from town (case ending) or town people (adjective). The main property of adjectives in Burushaski which distinguishes them from nouns is that they may be used with different classes. Quite a number of adjectives have different forms for singular, hx plural and y plural, much along the same lines as nouns: mat m ha r - a black horse (x singular) mat miu ha ra - black horses (x plural) s mat ming r ing - his black hands (y plural) and some are irregular: b rdum balt - a red apple (x singular) b rju balt - red apples (x plural) (the word balt = apple is irregular too, since it does not change its shape for the plural). But most have one form only; u = good is an example: s
                     

s u s u

hir - a good man (hm singular) b ltiu - good kinds of apples (x double plural) s

Comparison Comparison (more) is expressed with the location ending -cum = from in the following way: J cum n but hem b i. - He knows more than I do. me-from he much knowing is that is, From me as a point of reference he knows much.
 

Form j a j cum g gocum m mua c umucum s hara s harum ar num c d nyate h yatum g nccum gunc t yate
          

Ending -a -cum -cum -a -cum -a -um -um c -yate -yatum -cum

for me from me (lengthened vowel) from you (irregular, your-you-at-from) for her (with inserted -mu-) from my sister (with inserted -mu-) to town from town out of my hand (my-hand-in-from) on a stone / with a stone down from the house (house-on-from) from day to day (day-at-from day)

-yate - then, next (that(y)-with) Figure 7 Examples of location endings

-9-

For the superlative (most) comparing to khul = all is used: kh lcum u ha r - the best horse s all-from good horse that is, from all, the good horse. And the English expression the very best is rendered as the good of good, quite reasonably: gat nc - the very best trousers This expression looks remarkably like summa summorum, the book of books, etc. Adverbs There are very few adverbs in Burushaski, and those that exist are not derived from the corresponding adjective: hum lkum - quick s u - quickly An unexpected source of adverbs are the composite verbs (explained below):
       

s u cum u s


as p as p
 

t- - to sweep the floor (lit. to do as p) - swiftly

Adverbs of place The adverbs of place in Burushaski come in three series, one derived from akh - for close proximity, one from t - for further away and one from t- for at the other side; we show here the basic forms plus those with the simple direction endings, but there are many more:
 

Strange Things about Verbs The bad news is that Burushaski has perhaps as many paradigms as Latin, but the good news is that they are much more regular. Except for the verbs to be, to come and to go, there are no really irregular verbs. Therefore the paradigms can be given using any verb. Traditionally, (that is, both by Lorimer and Berger), verb a paradigms are shown for the verb -t- = to do. Since it requires an experiencer (a possessor), the actual form used is t- = to do it (y). Before we can present a paradigm, we first have to discuss two features of the Burushaski verb its three stems and its prefix and suffix scheme. After these and the paradigms, we will discuss the d-verbs, the consecutive, the imperative and negation. The three stems The Burushaski verb has three stems a past stem, a present stem, and a consecutive stem. The past stem is the unadorned root of the verb, and it is the one given in the dictionary. It is used for the past tenses but also for other purposes, for instance for the imperative, and, with the directional ending -a = towards, for a kind of infinitive of purpose: ta = in order to do it. The present stem is in principle derived from the root (past stem) by suffixing -: bi - = to throw a y c s object has bi - for present stem. But if the root ends in a consonant, the -- modifies this consonant s c c according to a complicated set of rules; to avoid problems, the dictionary gives the present stem for each verb in addition to the past stem. In the verb t- = to do it (y), the t is absorbed into the --, and the present stem is c c -. The consecutive stem (used for such consecutive phrases as He picked up his coat and ...) is equal to the root, except that its initial consonant is hardened, as described in the section on phonetic features, and that it has the stress on the first syllable, regardless of where it is in the root. The consecutive stem of bi - = to throw a s y object is thus p a-, with hardening and shifted stress, and that of t- = to do it (y) is an unchanged ts since it has neither an initial consonant nor shiftable stress. Prefixes and suffixes Whereas active languages distinguish subject and object and ergative languages have actors and experiencers, Burushaski has subjects, actors and experiencers. Whether this hybrid system is original to Burushaski or the result of external influences on an originally ergative system, is unknown. As we have seen, the experiencer is the subject of the intransitive verb and the object of the transitive verb; the actor is the subject of the transitive verb. In I fell, I is the experiencer/subject; in I saw her, I is the actor/subject and her is the
         

akh - here akh la - hither akh lum - from here

to - there t la - thither t lum - from there

it - over there ta - to over there tum - from over there

- 10 -

experiencer/object. This takes some getting used to. Almost all verb forms have an experiencer prefix and a subject suffix. The experiencer prefixes are the same as the possessive prefixes; to emphasize this fact, we shall translate them as possessive in the literal translations. In ja a-w l-a = I (abs.) my-fell-I = I fell, the prefix a- designates me as the experiencer and the suffix -a designates me at the same time as the subject. In j mu-y c-a = I (obl.) her-saw-I = I saw her, the prefix mu- designates her as the experiencer and the suffix -a designates me as the subject, as in the previous example. Also, ja in the first sentence is in the absolutive, since it is the experiencer, and j in the second is in the oblique, since it is the actor. To summarize, the full intransitive verb form consists of experiencer-prefix experiencer-prefix

+ verb + subject-suffix (subject=experiencer) + verb + subject-suffix (subject=actor)

and the full transitive verb form consists of The forms are structurally the same, but the interpretation of the subject suffix differs. Note that the - in t- = to do it (y) is the experiencer prefix of class y. With all this out of the way, we can now turn to the actual paradigms. The paradigms To show the subject endings of the verb, we present here the paradigm for the Past I tense, which, unsurprisingly, is based on the past stem:
 "!  "  "  

Where there is a Past I tense, the Past II tense cannot be far. Its endings are formed by inserting an -m- in the previous endings:
 "!  "  " 

The Past II emphasizes the fact that it happened more than the happening itself. The translation with indeed gives the idea, but is too heavy. Note that c = they (x) has timi rather than temen. Almost the same endings are used for the Past III tense, in which the (past) stem is immediately followed by - sc- (pronounced - sts-!):
   "!  "    " 

Again the x plural form is unexpected. This is an example of a long unstressed vowel, since the stress remains on the verb: t-. One can easily imagine that t scam = I did it is composed of ta scam in which ta is the infinitive to do it and scam is the first person singular of some auxiliary verb sc- not used otherwise. The Past III is used for side issues, side lines in the story: Te zaman ule ut nc but qaim t bi n. that time-at camels(x) very expensive were(x) (Past I) = At that time camels were very expensive. Te zaman ule ut nc but qaim t bi n scimi, ... that time-at camels(x) very expensive were-as-you-should-know(x) (Past III), ... = Now, at that time camels were very expensive, ... where the first is part of the main story and the second is side information. The same Past II endings with the present stem provide the future (rather than the present) tense:
                

 "

 "

 "

 "

 "

 "

 "

 "

j un n m s t

t t t t t t

scam - I did it scuma - you scimi - he scumu - she scimi - it (x) scimi - it (y)

m m w w c k

t t t t t t

scemen - we did it scumen - you all scumen - they (hm) scumen - they (hf) scimien - they(x) (!) scimi - they (y)

 "

 "

 "

 "

 "

 "

 "

 "

j un n m s t

tam - I did indeed do it tuma - you timi - he tumu - she timi - it (x) timi - it (y)

 "

 "

 "

 "

 "

 "

 "

 "

j un n m s t

ta - I did it ta - you ti - he tu - she ti - it (x) ti - it (y)

m m w w c k

ten - we did it ten - you all ten - they (hm) ten - they (hf) ten - they (x) ti - they (y)

m m w w c k

temen - we did indeed do it temen - you all temen - they (hm) temen - they (hf) timi - they (x) (!) timi - they (y)

- 11 -

Again note an unexpected form, this time for mi = we. Now one would expect the fourth combination, the Past I endings with the present stem, to produce a fourth tense the present or a second future perhaps but this combination does not exist. The above paradigms may give the impression that the verb is t-, but we have to keep in mind that the a verb is actually -t- = to do and that the - = he/it (hmxy) is the experiencer prefix. When the experiencer (object) is not hmxy, the prefix changes and we seem to get a different verb, but that is just our outlook: n ja phut ti he(actor) me glance my-did-he n mo phut m ti m ne phut tu m mi phut m ten etc.
      

- he looked at me - he looked at her - she looked at him - you all looked at us

And when the experiencer is the subject, as with intransitive verbs, the prefix and the ending move in unison, as seen in the Past I tense of the verb -m n- = to become:
   

Compound tenses The above are the simple tenses, simple in that the verb forms consist of one word. The compound tenses are formed by using the past or present participle with forms of the verb ba- = to be. The past participle is derived from the past stem by adding -m after vowels (bi m = having thrown a y object), and -um after consonants, s except that the participle ending for the first persons singular and plural is -am: tam = having done it (I or we), tum = having done it (all other persons). The similarity to the -um endings of many adjectives may be significant, but unlike these adjectives the participle does not have plural forms, and is also different in other respects. The present participle is derived in the same way from the present stem: bi m = throwing a y c object), am = doing it (I or we), um = doing it (all other persons). c c There are many combinations of these participles with tenses of ba- = to be, of which we show here the present:
 "!  "  " 

Again the unexpected forms are marked with an exclamation mark. The verb ba- = to be exists in the present tense only (if the above forms indeed represent the present tense of ba-); in addition there is a past participle, which is used alone and in combination with other participles to express compound past tenses. Its form is normally bam = having been (...), but differs for hf (bum), x (bim), y (dul m), x plural (bi m) and y plural (bic m). Examples are ja hur tam bam - I had been sitting I having-been-sitting-I was mo hur tum bum - she had been sitting she having-been-sitting(hf) was(hf) Note the interesting and quite logical difference in meaning between
    

 "

 "

 "

 "

 "

 "

 "

 "

j un n m s t

c am c um c um c um c um c um

ba - I am doing it ba - you b i - he (!) bu - she bi - it (x) du - it (y) (!)

m m w w c k

c am c um c um c um c um c um

ban - we are doing it ban - you all ban - they (hm) ban - they (hf) bi n - they (x) (!) bic - they (y) (!)

ja un ne mo se te

am na - I became gum na - you m ni - he mum nu - she m ni - it (x) m ni - it (y)

mi ma we we ce ke

mim nen - we became mam nen - you all um nen - they (hm) um nen - they (hf) um nen - they (x) m ni - they (y)

 "!  "  "  "  "  "

 "   "  "  "  "

j un n m s t

c am - I will do it c uma - you c imi - he c umu - she c imi - it (x) c imi - it (y)

m m w w c k

c an - we will do it (!) c emen - you all c emen - they (hm) c emen - they (hf) c imi - they (x) c imi - they (y)

- 12 -

tum b i - he has done it (and now its finished) it-having-done he-is and c um bam - he was doing it it-doing he-was (while something else was going on) The present participle is considered by Burushaski speakers as the verb itself: asked for the Burushaski translation of to fall, the speaker will answer wal m = falling. (It will be clear that the dictionary form -wal c cannot serve since it is a grammatical abstraction, not a word in the language.) More Kinds of Verbs The above straightforward simple verb is actually a minority in Burushaski; there are two other kinds of verbs, compound verbs (not to be confused with compound tenses) and d-verbs, which together are more numerous. Compound verbs Although many verbs consist of a root, possibly with a prefix, at least an equal number of verbs consist of a word followed by an auxiliary verb, generally t- = to do it(y). We have already seen three examples: as p t- - to sweep the floor a phut -t- - to look at c p t- to type (on a typewriter) but many more exist. These forms are comparable in form and usage to the Japanese suru verbs: benkyoo suru = study do = to study. If the verb is transitive, very often a corresponding intransitive verb exists in which the t- = to do it(y) has been replaced by -m n- = to become:
a phat -t- to allow, to abandon phat -m n- - to get loose, to be left behind
            

Sometimes the word in front of the auxiliary verb is clearly a noun, as in and the word phat = left-overs (y) may perhaps explain
a phat -t- - to allow, to abandon But in many cases the word only occurs in that specific verb and it is unclear what part of speech it is, as is the case with as p, phut and p above. as p occurs also as an adverb with a vaguely related meaning c (swiftly), as we have seen, and I have occasionally translated phut by glance, but no such word exists. Morin and Tiffou [lit.ref. 3] relate p to the English word type, but it may also be onomatopoeic. c
 

ph ing


t- - to lie (lit.: to do lies, untruths)

D-verbs The above forms are those of about half the Burushaski verbs; the other half adds a d- before its verb forms. An example is dwas- = to stay behind:  

We see that the experiencer prefix has now turned into an infix, and a copy of the vowel of the prefix is inserted if the d- would otherwise collide with a consonant. The d- causes hardening of the consonant in the prefix, as we can see from: un duk wasa - you stayed behind The same insertion of the experiencer prefix occurs in the participles: mo d-um -wa-um bum s - she was staying behind she d-her-staybehind-ing was-she (this is from the present stem dwa- = to stay behind). As an exception, most d-verbs that have pre/infixes - s have a special form for the y experiencer, in which the infix is -u and stress falls on the root: 

ja ne mo mi

d wasa d wasi dum wasu dim wasen

I stayed behind he stayed behind she stayed behind we stayed behind

- 13 -

te ndes duw si - the gold (y) remained Just as non-d verbs, there exist d-verbs without experiencer prefixes, and with Type I, II and III prefixes. When the verb does not carry an prefix, generally a vowel is added before the d- is prefixed, as in the verb do q- = s to slacken: se than u do qi - the rope (x) slackened s In this case we know that the do- is the d- prefix + vowel, since the word for loose, slack is qum. s The various kinds of verbs reconsidered As we have seen, there are d-verbs and non-d verbs (2 kinds), and there are verbs without experiencer prefixes and with Type I, II and III prefixes (4 kinds). All eight combinations exist, and no combination is rare or special. So it would be nice if we could pinpoint what each of these features means, but that is easier said than done. More d-verbs than non-d verbs have a stative (describing a state) or passive meaning. More Type II verbs than Type I verbs have a transitive or causative meaning. And many Type III verbs involve both a direct and an indirect object. But these are general observations and no more than that. A Some roots occur in one verb only, for instance d-a- = to itch, as in - my foot itched ah tis d ami my-foot d-to-me-my-itched but many occur in clusters, of which a typical example is: c a rum - cold daur- - to get cold (hx) -c du ur- - to get cold (y) c ac d-aur- - to make cold, to cool Many real causatives have an -s- infix, which causes hardening:
A X-cum -mal- - to be afraid of X A smal- to frighten g r- to have a colour a skir - to dye A clear example of the use of the latter set is given by Morin and Tiffou [lit.ref. 3] (page 24):
         

Gat nc mal m g rum bi n Trousers (x plural) black colour-having are (x) = The trousers are black Mo g se gat nc mal m osk rum bu The woman (hf) trousers (x) black their(x)-dying is (hf) = The woman dyes the trousers black [It may be significant that for Proto-Sino-Tibetan a system has been reconstructed in which d-, b- and ware passive and stative prefixes and s- is a causative prefix; and then it may not.] The situation with Type III verbs is perhaps a little clearer. Its use almost always involves at least an indirect object: men A -men          

n mel men - he drank wine n m mua mel m meni - he gave her wine to drink he(actor) her-to wine her-caused-to-drink-he Here, the prefix m - may be a double prefix, consisting of mo- = her and a one-vowel prefix for the experiencer, mel = wine. In fact, such double prefixes have occasionally been reported for Hunza Burushaski ([lit.ref. 4], page 211): Langabrumo Maluo-mo muring-ete cil m etimi c c L. M. her-hands-on water her-its-did-he = he made L. pour water on M.s hands where m etimi can well be seen as a combination of mo- = her and timi = he did it. But all this remains speculation.
   

- to drink - to give to drink

- 14 -

Consecutive, Imperative and Negation To complete the picture of the Burushaski verb, three more forms have to be explained. The consecutive The next verb form to be discussed is the consecutive, which has no counterpart in English. It has the meaning of after having done so and so or when such and such state had arisen; it is a kind of adverbial past participle and it is used very, very frequently in Burushaski. If the verb has an experiencer prefix, this shows in the consecutive; otherwise the consecutive is the same for all persons. In d-verbs it is the past stem without any ending, but with a possible experiencer infix: se mu d aur ... - the fish (x) having got cold, ... c c te cel du ur ... - the water (y) having got cold, ... c For non-d verbs it is formed by prefixing nu to the past stem without any ending. The in the nu indicates that it causes a shift of the stress to the second syllable. This nu also causes hardening of the initial consonant of the root (or of the go- prefix): nuk a - having-gone-to-sleep c from gu - = to go to sleep (note the stress shift). c If the verb starts with a vowel or has an experiencer prefix, n- rather than nu- is prefixed to form the consecutive: n yec - after having seen me nuk yec - after having seen you num yec - after having seen her ... / he saw her and ...
a a from -y c- = to see. The consecutive of -t- = to do is n-t, which gets the experiencer prefix inserted. So we get
        

as p net - after having swept the floor, ... phut nat - having looked at me, ... phut nom t - having looked at her, ... A typical example of the extent to which the consecutive is used is the following sentence, taken (and simplified!) from [lit.ref. 1], story V: lji ni te bult nuk a; c kunc t di ya tu q-ce n ya dar a gal . = He went home and lay down to sleep that day; the next day he got up, took his rifle and went hunting. which contains four consecutives: lji ni - Inside-to having-gone (from gal = to go, completely irregular) te bult nuk a - that day having-gone-to-sleep c (from gu - = to go to sleep) c kunc t di - the-next-day having-got-up (from di - = to get up) ya tu q-ce n ya - his rifle having-taken his-own rifle-with having-provided-himself (from X-ce -y - = to provide oneself with X) dar a gal . - to-hunt went-he. This device allows the Burushaski speaker to squeeze in a lot of information, and Burushaski text is generally more compact than its English translation. (The next sentence in story V starts B rule ni ... = A valley-in having-gone ...) [Burushaski kunc t = the next day and gunc = day may be related to Yeniseian kans = yes terday.] The imperative As in many languages, the imperative singular is equal to the root of the verb:
                     "  # #  

- 15 -

as p et! - Sweep the floor! J a te ndes bi ! - Throw me the gold! s me-to the gold throw The plural is formed by adding -ing: Phut m ting! - Look at her! (to several people) glance her-do-you-all There is also an imperative for the third person, which ends in -is for the singular and in -isen for the plural hx; the -i is dropped after a vowel. Examples are: Til ay lis. - He should not forget it. forgetfulness not-his-make-sting-should-he We n sen. - They should go. A more polite imperative can be formed by adding the question particle -a after the form: as p t-a. - Could you please sweep the floor. and a polite invitation is expressed by the second person future: Goh m a m ltiruma! c - Do show us your house! your-house us-to our-show-will-you
A from -ltir- = to show.
         

Negation There is no word for not, although there is be = no and aw = yes. [Yeniseian also has be = no.] A sentence is negated by prefixing the verb form with a = not (ay before vowels). Just as the consecutive prefix nu, the negative prefix a moves the stress to the second syllable and causes hardening of the following consonant; also, if the verb has an experiencer prefix, the i- which is normally left out, is kept:
  

In compound forms, the entire form is negated rather than the final verb: guw lum ba - you have fallen ak walum ba - you have not fallen A form in which the final verb is negated exists, but has a completely different and idiomatic meaning: am tuk guw lum ap - this way you wont fall now your-having-falling not-is in which guw lum = your-having-fallen acts as a noun as it does in English and ap = not-is-it(y) is a y form referring to the noun. (Due to the irregularities of the verb ba- = to be, the negative of du = it (y) is (itself irregular), is ap = it (y) is not, which is clearly derived from the negation a and bi = he/it (hmx) is through hardening.) The hardening of the consonant after a also extends to the d- of d-verbs: un duk wasa - you stayed behind un at kuwasa - you did not stay behind in which both the -t- and the -k- result from hardening. The consecutive cannot be negated: after the water had not got cold, ... is not a reasonable thing to say. [Yeniseian has at = not; if it is related, the final -t may even explain the hardening that the Burushaski a causes.] Noun forms from verbs By putting the ending -as after the past stem of a verb V, a noun or adjective is created that indicates someone who habitually or professionally does V. An example of the latter is c cenas = carpenter from c cen- = to work wood.
            

ja ja un un ne ne

aw la ay wala guw la ak wala w la a wala

I fell (from -w l- = to fall) I did not fall (ay before vowel) you fell you did not fall (hardening and stress shift) he fell (i- suppressed) he did not fall (i- restored)

- 16 -

Syntax and Subordinate Clauses The normal Burushaski sentence starts with zero or more consecutives, and ends with perhaps a subject (actor), perhaps an object (experiencer), and always a verb. The example given under The consecutive is typical. Although the texts seem to avoid having both an actor and an experiencer, examples can still be found easily: ne saudag re mo gus muc m bam the merchant (actor) the woman (exp.) her-carrying-away was-he = the merchant married the woman (from -c - = to take (away)). The order experiencer, actor, verb also occurs; this seems to have a more passive meaning: ne hir c le c mi the man (experiencer) water (actor) him-carried-away-it = the man was carried away by the water but we should not forget that the ending -mi in c mi refers to c le = water (actor) and not to ne hir = the man (experiencer). To have As in many languages, there is no word for to have: I have money is expressed as my money is, (as it also is, for instance, in Hungarian, where we have p nzem van = money-my is): j pais du - I have money of-me money(y) is(y) and of course the more probable - I have no money j pais ap of-me money(y) not-is(y) Subordinate clauses The constant use of the consecutive reduces the need for other forms of subordinate clauses. A number of subordinate relations can be expressed by using the past or present participle with location endings, in ways which can often be translated literally into English: G lta c rumyate - when he came to Gilgit Gilgit-to his-coming-on (upon coming to Gilgit) A word that features often in subordinate relations is ka; in isolation it means still, yet, and then, but it is also used to list nouns: huk ka bu s - a dog and a cat huk ka bu tul - a dog, a cat and a snake s dog and-then cat snake First it is used like the English that to connect two sentences: an ka se bu w lum bi s he-saw and-then the cat its-having-fallen is = he saw that the cat had fallen Its second usage is for conditionals: Ph ing tum ba ka un gus cam. lies having-done you-are and-then you your-eat-will-I = If you have lied, I will eat you. (lion to fox) (It is clear that the meaning cannot be If I have lied ..., since that would be Ph ing tAm ba ka.) Its third usage is for counterfactual conditions, where it is reinforced by the particle ceq, which seems to mean something like but not really and which absorbs the forms of the verb ba- = to be: Ph ing tum ceq ka un gus cam ceq. lies having-done but-not and-then you your-eat-will-I but-not = If you had lied, I would have eaten you. Occasionally, a past participle is used as an adjective, in a construction that corresponds to a relative clause in English:
                  $       

- 17 -

j girm num gut ket p - this book that I have written my its-having-written this book but in general relative clauses play a far smaller role in Burushaski than in English. Questions Yes-no questions are formed by suffixing -a after the verb form: G pais du -a? - Do you have money? of-you money it-is eh? Many question words agree with the gender of the noun they refer to:
         

men men mes met

d las das n bu s ha

which boy (hm) which girl (hf) which cat (x) which house (y)

Others are adjectives and end in -um: b ltum hir - what kind of man

Numerals The numbers one to three have different forms for the classes hmf, x and y; in addition they have a special form used, among others, for counting, the z class. Higher numbers have two forms only, one for the hxy classes, and one for counting. See Figure 8. The forms for 80 and 90 are reminiscent of the the French quatre-vingt = fourtwenty and quatre-vingt dix = four-twenty ten. The ordinal numbers are formed by adding the adjective ending -um to the form used for counting, except for the word for first, which, as in English, is not derived from the word for one. See Figure 9.

    

Figure 8 Numerals in Yasin Burushaski

100 1000 1999

tha haz r haz r hut

tha w lte- ltar turma-hut

19 20 30 40 50 60 70 80 90

11 12

turma-h n/-h n/-h k turma-alt n/-alt c/-alt etc. turma-huc /-hut ltar ltar t rum twenty ten alt - ltar two alt - ltar t rum two twenty ten isk - ltar three twenty isk - ltar t rum three twenty ten w lte- ltar four twenty w lte- ltar t rum four twenty ten

    

 

 

   

 

Class: 1 2 3 4 5 6 7 8 9 10

hmf hen alt n isk w ltu cend bi ndu s thal alt mbu huc t rum

Yasin Numerals x y han han alt c alt isk isk

z hek alt isk w lte cend bi nde s thal alt mbe hut t rum

- 18 -

Hunza Burushaski and Yasin Burushaski Compared One of the differences between Hunza Burushaski and Yasin Burushaski lies in one of the locative endings: where Yasin Burushaski has -a = to, towards, for (-a after vowels), Hunza Burushaski has -ar (-r after vowels). A few comparisons:


Another difference is that Hunza Burushaski uses go- = you, your as a possessive prefix only; all other forms derive from ung- = you. As we have seen, Yasin Burushaski uses un = you for the experiencer and the actor only; all other forms derive from go-. Where these two effects combine, that is in the form for to you, we get two completely different words, as marked in the examples above. A third difference is that third person masculine Yasin forms like tum b i = he has done it and um c b i = he is doing it contract to tai and ai in Hunza Burushaski. Also, other contractions occur that c are absent in Yasin Burushaski. The Hunza Burushaski sentence Langabrumo Maluo-mo muring-ete cil m etimi c c shown above would be Langabrumo Maluu-mu mur ing-yate cel m timi c c in Yasin Burushaski. The numerals from one to ten in Hunza Burushaski are shown in Figure 10; comparison with Figure 8 shows that the differences are minor. Hunza Numerals hmf x y hin han han ltan ltac lto sken sko sko w lto cund mi ndo s tal alt mbo hun c t rumo
             &          

References 1. Berger, Hermann, Das Yasin-Burushaski (Werchikwar): Grammatik, Texte, W rterbuch, Otto Harrassowitz, Wiesbaden, 1974, 228 pp. Although the book has the same structure as the three-volume work by Lorimer on Hunza Burushaski, the grammar is a real grammar, with rules and text in phonemic writing. The book is far from a text book,
'

&  

Figure 10 Numerals in Hunza Burushaski

Yasin h ra g smua j a g a m a
    

Hunza h rar g smur jar ngar(!) m mar

% 

Figure 9 Ordinals

to a man to a woman to me to you to us

Class: 1 2 3 4 5 6 7 8 9 10

haw lum alt um isk um w lteum cend um bi ndeum s thal um alt mbeum hut um t rumum
        

first (from haw l = the one at the front) second third fourth fifth sixth seventh eighth ninth tenth

z hik lto ski w lti cind mi ndi s tal alt mbi hunt t rimi

- 19 -

though; the material is presented in a linguistic rather than a pedagogical order, and requires several readings. The Burushaski texts contain some amusing and some X-rated stories. The dictionary is Burushaski-German, with a German-Burushaski index.
( ( ) (

2.

Tiffou, tienne, Jurgen Pesot, Contes du Yasin Introduction au bourouchaski du Yasin, tudes bourouchaski 1, SELAF #303, Peeters/SELAF, Paris, 1989, 163 pp. Assumes that the reader has read Bergers book, and extends it with more and new material. Paradigms given in full, more about vowel lengths, more stories. The dictionary is a word list to the stories only. Morin, Yves Charles, tienne Tiffou, Dictionnaire compl mentaire du bourouchaski du Yasin, tudes bourouchaski 2, SELAF #304, Peeters/SELAF, Paris, 1989, 58 pp. Additions and corrections to the dictionaries of Berger, Tiffou and Lorimer. Extensive terminology on weddings and marital customs. Lorimer, D.L.R., The Burushaski Language Vol. 1, 2, 3, Instituttet for Sammenlignende Kulturforskning, Oslo, 1935, 464 pp. Part 1 contains large amounts of very carefully described data, written phonetically rather than phonemically, organised according to subject. The book is not a grammar in that the author does not present rules but only reports observations. The author deliberately makes no attempt to distinguish between the results of the inherent inaccuracy of normal speech and those of grammatical processes and just recorded what he heard. Often the recorded material fluctuates so much that precise rules are difficult to find. Part 2 contains recorded and translated stories, part 3 is a dictionary. McCarry, John, Jonathan Blair, High road to Hunza, National Geographic, vol. 185, no. 3, p. 115-134, March 1994. Although the article concerns the Hunza valley rather than the Yasin valley, it is still very informative. Mook, John, Kimberley ONeil, Trekking in the Karakoram and Hindukush, Lonely Planet, Hawthorn, Australia, 1996, 322 pp. Contains short (about 50 words, 30 phrases and 20 numerals) vocabularies of Urdu, Balti, Burushaski, Kalasha, Khowar, Shina, and Wakhi (Xikwor). And of course descriptions of dozens of foot trips and tons of advice. Ruhlen, Merritt, A Guide to the Worlds Languages: Volume 1: Classification; with a postscript on recent developments, Edward Arnold, London, 1987; 1991, 463 pp. Blazek, V clav, D. Bengtson, Lexica Dene-Caucasica, Central Asiatic J., vol. 39, p. 11-50,161-164, 1995. A set of 219 etymologies linking Basque, Caucasian, Burushaski, Yeniseian, Sino-Tibetan and Na-Den (Athabaskan + Tlingit + Haida). The etymologies generally involve 3 to 4 members of the groups mentioned. A table of sounds correspondences is given. The literature references are copious.
) $ (

3.

4.

5.

6.

7.

8.