| Search Turner Nepali | About the Dictionary | Front Matter | Copyright Information | Back to DDSA |
[page xii]
INTRODUCTION
NEPĀLI,1 now the chief language of literature, administration, and general intercourse in the Kingdom of Nepal, has from time to time received various other appellations. These are (1) Gorkhāli, as the language of the Gorkhās (or Gūrkhās) who from their seat about the town of Gorkhā, under their king Prithwi Narain Sah (Pr̥thivī Nārāyaṇ Sāh) in the year 1769, finally made themselves masters of the Valley of Nepal and the present capital Kathmandu; (2) Khas-kurā or 'language of the Khas', to whom may have been due the introduction of this particular form of Indo-Aryan;2 (3) Parbatiyā or 'mountain-language', as opposed to the Newārī of the Valley or the Aryan languages spoken on the Plains of Hindustan.
Nepali belongs to the Indo-Aryan family, that is to say, like most of the other languages of Northern India and of certain outlying parts, 3 it is derived from a form of speech, of which our earliest document is the R̥gveda. This language or group of closely related dialects, which it is convenient to call Sanskrit (a name more strictly speaking proper to the form in which it was later stereotyped and employed for more than a thousand years as the chief literary medium of the whole of India), was brought into that sub-continent by the Aryans, probably during the latter half of the second millennium B.C.
Sanskrit in its turn forms part of a larger group, called Indo-Iranian or Aryan. In this the other member, Iranian, includes among its most ancient forms the language of the Zend Avesta and the Old Persian of the Inscriptions of Darius, and to-day is spoken under different forms from the Caucasus to the Western borders of India, where it appears as Pashto and Baluchi.
Lastly Indo-Iranian was itself derived from an earlier form of speech, which, known to us only from comparison of its descendants, is commonly termed Indo-European or Indo-Germanic. From this Indo-European language are descended also the ancient Tocharian of Chinese Turkestan, the Slavonic and Baltic languages, Armenian, Greek, Italic (with Latin and its modern Romance descendants), Keltic, and Germanic; to which is probably now to be added the ancient pseudo-Hittite.
The proof that Nepali is descended from Sanskrit rests upon the fact that many details of its grammatical structure find their explanation only in the corresponding forms of the earlier language, and that much of its vocabulary, allowing for a regular correspondence of sounds between the two languages, is identical with that of Sanskrit. Thus the opposition between the 2nd singular hos 'thou art' and the 3rd singular ho 'he is' (earlier hoi, as in the negative hoi-na) has its counterpart in Sanskrit bhávasi: bhávati (which became in Prakrit hosi: hoi); or that between 3rd singular ho and 3rd plural hun in Sk. bhávati: bhávanti (Pk. hoi: honti). The present participle jã ̄do 'going' corresponds with the stem of the Sk. present participle yāˊnt-, the past gayo 'went' with Sk. gatá-. In the adjective the masculine termination -o: feminine -i corresponds with Sk. -ako: -ikā (Pk. -ao: -iā). In the pronoun the nominatives jo and ko have oblique forms jas and kas: in Sanskrit these are yó and kó: yásya and kásya. Even more striking is the opposition in the correlative pronoun between nominative só and oblique tas which corresponds exactly with that of Sk. só: tásya.
The etymological notes which follow in the dictionary show that about five thousand of the words there collected are descended from corresponding words in Sanskrit. These words display a quite regular correspondence between the sounds of Nepali and the sounds of Sanskrit: thus words beginning with gh in Sanskrit will, if they survive, begin with gh in Nepali; Sk. ṣṭ = N. ṭh, Sk. rm = N. m, Sk. intervocalic -ṭ- = N. -r-, etc., etc.4
_________________
1Sir G. A. Grierson uses the Sanskrit form Naipālī in LSI.
2See Lévi, Le Népal, p. 259; LSI., vol. ix. pt. iv, p. 2 ff.
3The chief of these are Romani (i.e. the Gypsy dialects of Europe, Armenia, and Asia), the Dardic dialects of the Hindu Kush, Kashmiri, Western and Central Pahari, Eastern Pahari (Kumaoni and Nepali), Assamese, Bengali, Oriya, Bihari, Hindi, Panjabi, Lahnda, Sindhi, Gujarati, Marathi, Singhalese.
4Examples: Sk. ghr̥tá-: ghiu, gharmá-: ghām; kāṣṭhá-: kāṭh, piṣṭá-: piṭho; kárman-: kām, cárman-: cām; kīṭá-: kiro, caṭaka-: caro, etc., etc.
[page xiii]
Whereas the derivation of Nepali from Sanskrit cannot be in dispute, its exact position within the Indo-Aryan family is more open to discussion. With which of the modern Indo-Aryan languages is Nepali most closely allied? Indisputably its nearest relative is its western neighbour Kumaoni, a group of dialects spoken in the British Indian District of Kumaon. And in fact all the Indo-Aryan languages along the southern face of the Himalayas have certain features in common. This is intelligible if these languages were carried into their present habitats by the migration of the Khas (Sk. khaśa-) from their earlier home in the North-West.1
The most reliable linguistic evidence for the original close connection between two dialects of one language lies in the existence in both of common innovations not shared by other members of the whole group. The phonetic changes which differentiate the modern Indo-Aryan languages from Sanskrit are for the most part common to the great majority of them. There is one, however, which is of importance for establishing the earlier connections of Nepali.
In the group nasal + unvoiced consonant, the majority of the Indo-Aryan languages have preserved the consonant unchanged: thus Sk. dánta- becomes Assamese, Bengali, Hindi, Gujarati dã ̄t, Oriya dānt, Marathi dāt, Singhalese data. But in one group, that of the North-West, the consonant has been voiced: thus Kashmiri dand, Panjabi dand, Sindhi ḍ ̠andu. This change is shared by nearly all the Pahari languages, and runs into Nepali, e.g. dã ̄de 'harrow', kāmnu 'to tremble' from *kã ̄bnu from Sk. kámpate, kã ̄ṛo 'thorn' from Sk. káṇṭaka-.
At the time when this phonetic change took place we cannot say where exactly those who spoke the dialect which was to become Nepali were situated; but probably they were far to the west of their present home. For the change is comparatively ancient, since it has affected the Gypsy languages.2 The original location of the Khaśas in the North-West would therefore explain their sharing in this change.
Of the other phonetic innovations important for determining early dialectal connections there is not much to be found in Nepali. In its treatment of Sk. kṣ > kh it agrees with Gujarati, Sindhi, Lahnda, Panjabi, Hindi, and the Eastern group: contrary examples with ch < kṣ, as churi 'knife', chār 'ashes' from Sk. kṣurá-, kṣārá- are found as loans with ch in all these languages. In its treatment of the t of r̥t as a dental (subsequently disappearing) it agrees with the same group (except perhaps the Eastern languages): for of the contrary examples, moro 'corpse' (< *maro < maṭaka- < mr̥tá-) occurs in this specialized sense with a cerebral as early as Pali maṭaka- 'corpse' beside mata- 'dead', and māṭo 'earth' < mŕ̥ttikā is found everywhere with a cerebral, except in Marathi and one dialect of West Pahari.
In its treatment of Sk. -īya- > Middle Indian -īa- (as opposed to -ijja-), as in the passive suffix -i-, it agrees with Sindhi, Lahnda, Gujarati, and Hindi.
Nepali appears, then, to have belonged originally to a dialect-group which included the ancestors of Gujarati, Sindhi, Lahnda, Panjabi, and Hindi. In one particular it was closely associated with the most northern and western of these, namely Sindhi, Lahnda, and Panjabi. It is differentiated from the Dardic group (in which kṣ > c̣c̣h and probably r̥t > aṭ or īya > ijja), from Rajasthani (Marwari -īj- < -īya-); from Marathi (in which kṣ > ch > s, and r̥ a, and -īya- > -ijja-); from the southern group, Ardhamagadhi Prakrit and Singhalese (in which r̥t > aṭ); and perhaps from the Eastern, Bengali etc. (in which rt > ṭṭ).
The close resemblance, noted by Grierson,3 of the Pahari languages with the Rajasthani is due rather to the preservation of common original features than to the introduction of common innovations.
But once the speakers of these languages which were carried eastwards along the foothills of the Himalayas had settled in their new homes and had lost touch with their linguistic relatives of the North-West Panjab, their subsequent development was independent. Being brought into close contact with the dialects of the Plains to the south, they shared with them subsequent important sound-changes. Thus whereas in the Panjab the Middle Indian group, short vowel + double consonant, remained unchanged, further east and south the consonant was shortened and the vowel lengthened: Sk. mattá- became Panjabi mattā, but Hindi mātā. This change the ancestor of Nepali now shared (N. māt).
Whereas in the North-West and West (Lahnda, Sindhi, Gujarati, Marathi) MI. -ṇ- (< Sk. -n- or -ṇ-) remained, in the Centre and East it became -n-: so too in Nepali (though not in the Pahari languages to its west).
_________________
1LSI., vol. ix, pt. iv, p. 8.
2Turner, Position of Romani in Indo-Aryan, p. 23 ff.
3LSI., vol. ix, pt. iv, p. 2.
[page xiv]
Immediately to the south in East Hindi and Bihari dialects MI. -ḍ- and -ḍḍ- are distinguished as r and ṛ: so too in Nepali.1
The same influence was felt in grammatical innovations of a comparatively late date: thus the genitive affix ko is the same as the Hindi kā.
Despite the fact that Nepal has for long been an independent kingdom, this predominating influence of the language of the Plains has continued, and is even increasing in force. Hindustani, in vocabulary and idiom, is affecting the spoken, as its written form of High Hindi is affecting the written, language -- all the more that the majority of Nepali printed books are either translations from, or imitations of, Hindi works.
Whether when the speakers of Nepali arrived in their present country, there were already spoken there other Indo-Aryan dialects, is not known. But it is at least probable: for Indo-Aryan languages occupied the plains to the south from the time of Aśoka, and intercourse with the Valley, despite the obstacles of jungle and mountain, had been carried on. If there were such an Indo-Aryan language, it was probably closely akin to the ancestor of Bhojpuri and Maithili. The court language of Pāṭan about 1650 B.C. appears to have been of this type.2 We shall see later that a not inconsiderable number of Bihari words are to be found in Nepali.
Open as Nepali has been to the influences, for the most part cultural and literary, of the Indo-Aryan languages to the South, it has been still more exposed to the action of non-Aryan languages, and especially of those belonging to the Tibeto-Burman family. For a great many of those who learnt to speak Nepali had originally as their native tongues Tibeto-Burman or other languages.3
It is possible that Nepali has to some extent been influenced by this, and that certain idioms, as for example the honorific conjugation and the use of the agent postposition with the subject of transitive verbs even in the present, may have been, as suggested by Grierson,4 imported into the language by original Tibeto-Burman speakers. But such influence can be exaggerated; and most of these idioms have parallels elsewhere, and might be explicable within the language itself even if there were no question at all of foreign influence.
VOCABULARY
It is with the vocabulary that the dictionary is chiefly concerned. Six main streams have contributed to its formation.
1. At its core is the vocabulary it received, through the slow process of linguistic evolution, from the original Sanskrit or Indo-Aryan: the pronouns, the numerals, many adjectives and substantives, especially those expressing concrete ideas, and a large proportion of its verbs. But it by no means follows that, because a given Nepali word is descended from Sanskrit, it has come down, as it were, in the direct line of descent. From the time of the R̥gveda onwards there is ample evidence of the extensive inter-dialectal borrowing that has proceeded in the Indo-Aryan languages. If in Nepali we have kāṭnu 'to cut' < Sk. kártati and kātnu 'to spin' also from a *kartati (cf. Sk. kárttum, kr̥ṇátti), only one can be the legitimate Nepali form: the other (in this case probably kāṭnu) was at some period borrowed from a dialect in which Sk. rt became ṭṭ (and not tt). This must hold good of very large numbers of the Indo-Aryan words in Nepali (as in all the other Indo-Aryan languages); but it is only occasionally that we can glimpse the process. For in the case of most old loanwords from connected dialects any differentiation that existed at the time of borrowing has since been overlaid.
2. In addition to the words received eventually from the period of Indo-European or at least of Indo-Iranian unity, there is a large body of words, common to many of the Indo-Aryan languages, but not traceable to that earlier source. Some of these, of the onomatopoeic type, may have been new creations; but most, without doubt, have
_________________
1Festschrift Jacobi, p. 43.
2LSI., vol. ix, pt. iv, p. 17.
3It is possible that some of the native languages of Nepal are of Muṇḍā origin: see Les Langues du Monde, p. 399; LSI., vol. iii, pt. i, pp. 179 ff.; Turner, in The Gurkhas, p. 66.
4LSI., vol. ix, pt. iv, p. 15.
[page xv]
been imported from the languages with which Aryan came into contact. They include notably the names of plants and animals, and of articles hitherto unknown to the invaders. But there are also in this category words of more general import.Some of these words appear in the oldest Indo-Aryan texts, in the R̥gveda, and especially in a text of more popular character, the Atharvaveda. In subsequent texts, both of Sanskrit and its later forms of Pali and Prakrit, they occur in ever-increasing number.
It is now almost certain that the invading Aryans found established in India at least two other great families of languages, the Muṇḍā (belonging to a larger whole, the Austro-Asiatic) and the Dravidian. The Muṇḍā languages, which survive in certain isolated parts of India to-day,1 were perhaps at one time spread over the whole of Northern and Central India.2 The Dravidian lay to the West and South.
The researches of Przyluski have shown it to be not unlikely that Muṇḍā-speaking populations maintained themselves till a late date in the North-West of India,3 and that the origin of some at least of the non-Indo-European vocabulary of Indo-Aryan is to be sought in the Muṇḍā languages.4
The etymologies, particularly of the words referred to in the Index beginning on p. 657, reveal a certain phonetic peculiarity. They show, from language to language, an interchange of sounds -- dental with cerebral, aspirate with non-aspirate, voiced with unvoiced -- which cannot be accounted for by the normal correspondences of Indo-Aryan. This may indicate the borrowing of words either from languages with sounds unfamiliar to the borrowers -- such, for example, as the unvoiced mediae, which might well be reproduced either as unvoiced or as voiced -- or from different dialects of the same language, in which these correspondences were regular.
3. It has been suggested above that many of the words of Nepali, although derived from a Sanskrit source, have not come down in the direct line of descent, but must at one time or another have been incorporated from other Indo-Aryan dialects. In more recent times Nepali has borrowed a very great number of words from neighbouring Indo-Aryan languages.
The chief source for these has been, and still is, Hindustani. From it have come not only words of common Indo-Aryan origin (as, e.g., ghoṛā 'horse' from Sk. ghoṭaka-, which in Nepali would have had the form *ghoro, or paṛnu 'to read' from Sk. páṭhati, beside parnu, the form proper to Nepali, but confused with parnu 'to fall'); but also a very great number of words -- substantives, adjectives, and adverbs -- which Hindustani itself has received from Persian or from Arabic through the medium of Persian. One class in particular, that of legal terms, belongs here.
A smaller number of words in Nepali have forms proper to the East Hindi and Bihari dialects immediately to the South. Those which contain an original Sanskrit intervocalic -l- are most easily identified as loans of this sort. For Sk. -l- remains unchanged in Nepali, but becomes r in East Hindi and Bihari dialects. Thus hori beside holi, barad 'ox' < Sk. balivarda- (H. B. balad), hardi beside haldi, haris 'beam of plough' < Sk. halīṣā (L. haleh, M. haḷas), and in general a considerable number of agricultural terms.
4. The existence in the country of many Tibeto-Burman languages, one of which, Newari, produced a considerable literature in the Valley of Nepal, added to the fact that a large proportion of Nepali-speakers originally used these other languages, might at first sight make it probable that Nepali should have borrowed largely from the vocabulary of these languages (it has already been shown that the idiom was perhaps influenced by them). But in fact the number of words of Tibeto-Burman origin appears to be small, and to be confined almost entirely to names of specific objects. This is in accordance with general linguistic experience.5
5. A certain number of English words have been, and are being, adopted. They are brought in both by the literate classes either directly or through the medium of Hindustani, and by the soldiers recruited for the Gurkha regiments of the Indian Army.
_________________
1LSI., vol. iv, p. 7.
2S. Lévi, J A. 203, p. 56; J. Przyluski, J A. 208, p. 53.
3J A. 1926, p. 54.
4MSL. 1921, p. 208; BSL. 1922, p. 119, 1924, p. 69.
5Speaking at the Seventeenth International Congress of Orientalists in Oxford in 1928 Professor H. Lüders noted the observation he had made on Gurkha prisoners of war in Germany that, whereas Nepalese speakers of Tibeto-Burman languages used large numbers of Indo-Aryan (Nepali) words, the converse was not true.
[page xvi]
6. Finally, the greatest source of borrowing, as with all the modern Indo-Aryan languages spoken by Hindu populations, has been Sanskrit itself. Such borrowing has without doubt proceeded in these languages from the earliest times when Sanskrit had acquired prestige as the language of culture and was preserved as such in an archaic form, differentiated in ever increasing measure from the languages to which it had given birth. Some of the earlier borrowings of this sort have in their subsequent evolution lost all trace of their borrowing; and many words which as far as their sounds are concerned might be inherited in unbroken transmission, may in actual fact have been borrowed from Sanskrit. Even when the inherited word exists or existed, there has been a tendency to replace it with the equivalent Sanskrit word, to use which is a mark of culture. Such loanwords may appear either in their complete Sanskrit form (losing only in certain cases a final short -a of the Sanskrit stem) or, when widely used by all classes, may have been made to conform to the general phonetic system of Nepali. Thus the forms vidyā and bidde, or the words saṁtoṣaṇā and santok are equally loans from Sanskrit; but the former in each case is confined to learned circles, who may be presumed to have some knowledge of Sanskrit, while the latter is used by the people generally. In some cases, where phonetically both the borrowed and the inherited word would have the same form, it is difficult to make a decision. Are the words dos 'fault' and ban 'forest' inherited or borrowed from Sanskrit? Sometimes light may be gained from comparison with other modern Indo-Aryan languages. In this case, for example, the Sindhi forms (and, in one case, the meaning) -- ḍ ̠ohu 'fault' and vaṇu 'tree' -- show them to be inherited in that language.
The two main channels for the entry of borrowed Sanskrit words into Nepali have been literature and religion. Sanskrit is the one inexhaustible source for all words of a learned character. The translations of works like the Rāmāyaṇa and the Mahābhārata, which are filled with Sanskrit loanwords of this sort and are freely read and recited, have greatly helped to popularize large numbers of these borrowings.
In the etymological notes, which are included in square brackets at the end of the articles in the dictionary, the words borrowed from Sanskrit are clearly distinguished from those inherited or descended from Sanskrit.
A borrowed word is indicated always by the letters lw. (loanword) followed by the Sanskrit word from which it is borrowed.
For an inherited word first is given the Sanskrit word from which it is derived, then such Middle Indian forms (Pali, Prakrit, etc.) as can be ascertained, and lastly, following always the same order,1 cognate forms from the other modern Indo-Aryan languages. In some cases only a Middle Indian form can be given (and even when a Sanskrit word is quoted it may often be only a late sanskritization of a Middle Indian word). In other cases no ancient form at all may be available, but only cognate forms from other modern languages. With words of this class it has sometimes seemed advisable to reconstruct, from what is known of the phonological equations of Indo-Aryan, a hypothetical form which may carry the history of the word in question a little further back, and perhaps enable some other investigator to establish its origin from another source.
As has been indicated, many Nepali words are borrowed from Indo-Aryan languages other than Sanskrit, especially from Hindustani. When the borrowing is clear, in this case also it is indicated by the letters lw.; and if the Hindustani (or other) word from which it is borrowed is itself of Indo-Aryan origin, notes on its etymology follow as in the case of original Nepali words.
But if it, in its turn, has been borrowed from some non-Indo-Aryan source, usually that source is simply stated without further discussion of its original form. Thus under muluk the statement '[lw. H. mulk fr. Ar.]' indicates that this word has been borrowed from Hindustani mulk, which in its turn has been borrowed from Arabic. Such loanwords from Hindustani, which have their ultimate origin in Persian or Arabic, are very numerous.
ORTHOGRAPHY
The Devanāgarī alphabet, as finally developed for writing Sanskrit, was admirably phonetic. Its chief weakness was the principle by which, unless otherwise indicated, the vowel a was always understood with each consonant. This led to the clumsy and often intricate use of conjunct consonants. For a consonant could be written by itself, without
_________________
1This order will be found on p. 655 and is followed in the indexes.
[page xvii]
any following vowel, only if it were attached to a following consonant, or were marked with the virām (Sanskrit virāma-), an oblique stroke placed below the consonant-symbol. In practice the former method was followed wherever possible. Nevertheless, despite this drawback, Sanskrit was in this way written approximately as it was pronounced.As in other countries, earlier generations were less conservative in their spelling than those of to-day; and spelling, even if it lagged somewhat behind, was in general changed to meet changed pronunciation. For the most part the medieval forms of the Indo-Aryan languages were, when written in Devanāgarī, phonetically spelt. A Sanskrit word हस्तो hasto, which in course of time had become, say, in Hindi hātha, was spelt हाथ, or an earlier (Prakrit) sarantao, which had become saratā, was spelt सरता. But, this stage reached, spelling became much more conservative. One of the most important phonetic changes, differentiating the modern Indo-Aryan languages from the medieval, has been the loss of short vowels in the final position or in the interior (especially in this case short a) between consonants. Thus, in Hindi, hātha has become hāth in pronunciation, but it is still written हाथ, i.e. hātha, not हाथ्; saratā has become sartā, but it is still written सरता, i.e. saratā, not सर्ता or सर्ता.
Nepali, however, because it had not, like Hindi, a long literary tradition behind it, was less conservative. For more than a hundred years use has been made of the more simple conjunct consonants for writing real Nepali words: thus spellings like सर्दा, बोल्दा, भान्सा for sardā, boldā, bhānsā (from earlier saradā, boladā, bhānasā) are frequent. The virām was used both in the interior of words in avoiding awkward combinations of letters, and finally. Thus सन्कु or सक्नु saknu, छन् chan, हुन् hun, हवस् hawas, etc., etc.
In the final position, however, the use of the virām is generally confined to verbal forms. The reason is this: whereas very few verbs are borrowed from Sanskrit, innumerable nouns are. The Nepali verbs therefore have in writing been free from the clogging conservatism of Sanskrit spelling. On the other hand a noun borrowed from a Sanskrit noun ending in -a (such as sthāna-), although in Nepali the final -a is no longer pronounced, is still usually written as in Sanskrit, that is to say with final -a (i.e. स्थान, sthāna, not स्थान् sthān, as actually pronounced). This principle has been extended to inherited Nepali nouns; and generally, for example, हात (i.e. hāta) is written for hāt. This has the added inconvenience that both forms -- hāt and hāta -- which may have different signification (see, for example, s.v. hāt) exist, without being distinguished in the current form of writing.
The first to break free from this conservative tradition was A. Turnbull, who in his grammar employed the virām whenever a word, whether verb or noun, did in pronunciation end in a consonant; and he rightly used it not only with inherited but also with Sanskrit loanwords, when so pronounced, writing, e.g., स्थान् sthān, but अन्त anta.
This is the principle that I have adopted throughout the dictionary. In the middle of the word I have used the simplest forms of conjunct consonants1 and the virām where such are not available; at the end of the word the virām is always used.
It is this practice more than any other that the Sanskrit-educated reader will find repugnant. But let him consider that we are writing Nepali, not Sanskrit; and that Nepali is a language worthy to possess its own system of spelling.
As in Bengali, and perhaps in Gujarati, there is in spoken Nepali no longer any distinction of quantity or quality between long ī and short i, or between long ū and short u. i represents equally Sanskrit i, i before two or more consonants, and ī. So too with u. There is no difference in actual pronunciation between uni 'they' (from older unhi) and uni 'woollen' (from older ūnī). There is in consequence, as in Bengali and Gujarati, much hesitation in the use of the long and short forms in writing. Of late years there has been a certain tendency to write the short forms in the interior of words, the long when they are final. But there is no justification for such a practice. And since there is no distinction in pronunciation I have uniformly used the short forms.
The only exception to this has been in the case of words borrowed from Sanskrit which are, as far as I could ascertain, at present used only in learned circles. If the Sanskrit word, however, has been thoroughly adopted, I have used the short vowel, because, however shocking it may be to the Sanskrit scholar, it is actually so pronounced in Nepali. Thus git or gid and bhut are loanwords from Sk. gīta-, bhūta-.
Intervocalic -h-, being no longer generally pronounced, has been omitted in writing, except in certain cases in which, forming a syllabic boundary between two similar vowels, it is still pronounced.
_________________
1See the note at the end of the Introduction on p. xix.
[page xviii]
Similarly r and ṛ are written for rh and ṛh, in which the aspirate is no longer to be heard.
Under certain conditions, especially in the final syllable, an earlier yā has become e. Thus गनर्या garnyā has become garne, and is quite generally so written (or by Turnbull as ye). I have in such cases chosen the more
Similarly ya has become ε (open e). This is sometimes written ए e, but inaccurately. Since there is no symbol for ε, I have retained the spelling य ya, which always has this new value except initially. I have chosen therefore the spelling त्यस् tyas, not तेस् tes.
It is a frequent, if slipshod, practice to write anuswār (Sk. anusvāra-) instead of the appropriate nasal consonant in front of another consonant: thus अंत aṁta for अन्त anta. Here in all cases the nasal consonant has been written in full.
A vowel preceding or following a nasal is usually nasalized. This is sometimes indicated by writing the anunāsik (candrabindu) ँ or ¨ (and Turnbull followed this practice, writing दिंनु dĩnu, मां mã ̄, etc.); but it is often omitted. And since in this position the vowel is always nasalized, I have omitted it in writing: thus दिनु, म, मा, मास्, which can only be pronounced as dĩnu, mã, mã ̄, mã ̄s. In all other cases nasalization has been indicated by the anunāsik ँ written over the nasalized vowel. Only in the case of two vowels, either of which or both may apparently be nasalized, has it normally been written only over the second: thus आउँ āũ (which may also be pronounced ã ̄u or even ã ̄ũ).
It is probable that when a nasalized vowel is followed by a voiced consonant a full nasal consonant is developed after the vowel: thus dã ̄de is pronounced dã ̄nde, as opposed to dã ̄t which remains unchanged. In some words, e.g. sambhārnu, the nasal consonant has been written. But this knowledge came to me at a time when it was no longer possible to test the pronunciation by personal experiment. Consequently the dictionary will show some confusion in these spellings.
Greater difficulty was felt in deciding the spelling of words borrowed from Sanskrit (tatsamas). It has already been indicated that where the final -a is not pronounced, it is not in this dictionary written, the final consonant being given the virām. But popular pronunciation has considerably altered also the body of the word, especially in regard to certain combinations of consonants. Thus, for example, initial v- is pronounced b-, kṣ is pronounced ch or cch, ṣ (ष) is pronounced kh or k, jñ is pronounced gy or gg, r̥ is pronounced ri or ir, and so on. In these cases I have freely deserted the Sanskrit spelling, and adopted one phonetically representing the actual pronunciation of an ordinary educated man. Thus I spell bidyā, not vidyā; santok, not saṁtoṣ; chetri, not kṣatriya; āgyā, not ājñā; kirpā, t kr̥pā. Such spellings are frequent in Nepali works; nevertheless in a dictionary they may horrify the learned. Once again, let these reflect that this is a dictionary not of Sanskrit, but of Nepali. In the case of words still confined to purely learned circles, I have, as in the matter of long ī and ū, preserved the Sanskrit spelling, with the exception of adding the virām where necessary.
In most cases variant or learned spellings are included in the dictionary in their appropriate place, with references to that which is considered preferable and under which the meaning is given.
Lest I be accused of wanton innovation, let it be noted that no new spelling has been invented: only out of those varying systems already found in use has been chosen that which most nearly represents in writing the actual pronunciation of the spoken word. That is the aim of writing; and in so far as it moves away from that, it fails in its purpose.
The general adoption of the system of spelling advocated in this dictionary would, without any violent break with tradition, give Nepali a practical and phonetic system of spelling and would lessen the difficulties of teaching children and others to read and write.