Indian Scripts Index for Herbs and Spices

This index gives names for all spices in several languages of India. This includes all official Indian languages if written with a Brahmi-derived script (plus some from neigh­bouring regions); this means that both North-Indian (Aryan) and South-Indian (Dravidian) languages are accounted for. Those Indian languages written in Arabic script (Urdu, Kashmiri, Punjabi in Pakistan) are not included here, but will appear in the Arabic Index in due time.

The Brahmi-derived writing systems have a highly complex (but rather uniform) structure. They work basically alphabetically, i.e., they contain characters for both vowels and consonants; yet these are combined in a nontrivial and varied manner to yield symbols for syllables (a syllable here means a sequence of consonants followed by exactly one vowel). Inside the consonatic part, all consonants signs merge into complex ligatures called conjuncts. The vowel sign may be placed above or below the consonants like diacritics in the Latin alphabet; but some vowel signs follow or precede the consonant cluster (or in the case of two-part vowel signs even both at the same time). The vowel A s called implicit as it has no graphic representation of its own, unless it constitutes a full syllable without any preceding consonants (which is frequent word-initially, but extremely rare in other positions). Further complications arise from the large number of sounds and thus signs, and a small set of diacritical marks that modify the pronunciation.

The basic principles have different implementations in each of the modern scripts: Conjuncts may be formed horizontally, vertically or idiosyncratically, and in some scripts they involve subscripted characters; but Tamil has no conjuncts at all. In some scripts, the shape of a consonant depends on the vowel associated with it. Some scripts use two-part vowel signs one part of which precedes the associated consonant cluster (but in Sinhala, it only preceeds the last consonant of the cluster). And most scripts have one or few peculiarities not shared by the others (for example, the addak-sign in Gurmukhi that indicates double consonants, or the chillu shapes for word-final consonants in Malayalam). Also, some features common to all scripts may have different semantics (e.g., the anusvara or chandrabhindu signs may indicate nasalisation of the previous vowel or of the following consonant). Another puzzling case is the eyelash ra which appears in few Devanagari-written languages but whose precise meaning is not clear.

Each script is used for one or several languages, whose phonetics determine the character repertoire. Thus, one can usually not spell a language in an alien script without some dirty tricks. There is an exception, though: The North Indian Devanagari script was augmented with special characters and is close to a superset of all Indic scripts. These special characters are, however, rather an academic tool and hardly known to Indian speakers.

The romanizations given here are very close to both the common scientific transliteration (IAST) of Sanskrit and the enlarged ISO 15919 standard. Pure ASCII schemes like Harvard-Kyoto and ITRANS are very different.

• Retroflex consonats of Sanskrit are marked by a dot below the letter, (Ṭ, Ḍ, Ṣ, Ṇ). The same dot is used for retroflex variety of L (Ḷ). For R, however, I have to use two different special characters: North Indian ड़ (DDDHA, equivalent to DDA with Nukta) is rendered with a dot (Ṛ); its equivalent letter in Bengali (ড়) and Oriya (ଡ଼) is called RRA, not DDDHA. South Indian RRA ऱ (e.g., Tamil ற) is romanized with an underline, Ṟ.

• The Devanagari eyelash ra, which is characteristic of Nepali and Marathi, has no codepoint of its own; instead, it is coded as RA+VIRAMA+ZWJ र्‍ in Unicode 2.0, but Unicode 3.0 interprets the character as half form of RRA (and consequently writes RRA+VIRAMA ऱ्‍). For transliteration, I use Ř.

• Vocalic R and L are usually marked with a dot in scientific Sanskrit romanization (IAST). However, since the dot is already used for retroflex consonants in ISO-15919, I have to use a ring instead: R̥ and L̥. The long versions have an additional macron, R̥̄ and L̥̄. These characters might be poorly supported on some browsers (but fortunately, all except R̥ are rare or even nonexistent in Sanskrit, and do not appear at all in the modern languages except in learned Sanskrit loans).

• The velar and palatal nasals are written Ṅ and Ñ, respectively. Note that in most Indic alphabets, both are frequently written as anusvara.

• The implicit vowel is always rendered A, even if it is rather pronounced like O. Most North Indian languages often write this vowel even if it is not spoken, and in such cases my transliteration and sorting are often slightly inconsequent.

• Superscript H following a consonant (ʰ) marks aspirated sounds (which exist in all languages save Tamil and Sinhala).

• Superscript N preceding a voiced stop (ⁿ) indicates prenasalized stops (this feature is restricted to Sinhala).

• Long vowels (Ā, Ī, Ū) are distinguished by a macron. E and O are always long in Aryan languages, and thus not specifically marked (as in IAST, but not in ISO-15919). I write the Dravidian short E and O as Ĕ and Ŏ, respectively. There is also a short E and O in some North Indian languages to write English loanwords (transliterated Ê and Ô), but these do not appear in this index.

• The Gurmukhi and Devanagari letters KHHA and GHHA represent khah خ and ghain غ, respectively, in some Arabic loanwords. ISO-15919 suggests KH and Ġ as romanization. The former is in my opinion a poor choice, as it is not easily representable in Unicode. Thus, I use Ḫ and Ġ as in Arabic; however, neither of them appears in any spice name I could find. Also, the letters QA and ZA are used only for Arabic or Persian terms, and are consequently rare. FA, on the other side, appears in a few few native words plus Arabic, Persian and English loans.

• The Tamil letters LLLA ழ (also in Malayalam ഴ) and NNNA ன are represented by L and N with an underline, in accord with the scientific transliteration of Tamil: Ḻ and Ṉ, respectively.

Additionally, Devanagari transliterations are given for languages that have another native alphabet. These Devanagarizations use a couple of special signs to accommodate sounds alien to most Aryan languages. In general, there is a 1:1 correspondence between Devanagari and native letters; as an exception, the anusvara has been used in Devanagari throughout even if the native alphabet expresses nasalization by means of character, and a few native signs not reprensentable in Devanagari had to be replaced by near-matches. Since the Devanagari transliterations are produced programmatically by sed, some of them might by systematically flawed.

Displaying this page correctly is currently a true challange to any computer system, many of which will fail. You will not only need fonts for all scripts used, but you must also make sure that your browser (or the underlying operating system) can handle the complex rules of Indic typography correctly. Malayalam is an especially tricky case, and many current systems fail the following test: മഞ്ഞള്‍ (maññaḷ) and എള്ള്‌ (ĕḷḷə). You should see only one virama (a breve-like mark) over the last letter of the second word, while on some systems up to four viramas are shown, or characters are replaced with question marks or square boxes. Fortunately, Unicode 5.1 offers a new way to code the first word as മഞ്ഞൾ, which should be easier to interpret for renderers (as soon as the fonts are updated).

The entries are sorted according to the canonical Devanagari collating sequence, which is mimicked by all the other Indic scripts. Anusvara is sorted as if it were written as a nasal. The handling of the implicit vowel is still somewhat unsystematic; it is often ignored in sorting if it is written but not pronounced.

By its very construction, this index can hold only a limited number of scripts and languages; the following table gives an overview. This list is basically identical to the set of all Indian languages filtered by two criteria: (a) written in a Brahmi-derived alphabet and (b) having official or officially recognized status (either by the Union or by a single state); plus Sinhala. Since Bishnupriya Manipuri has at least an active community on the web, I have tentatively added it as a possible candidate for future inclusion. The 200+ more Indian languages are either written in Latin or Arabic alphabet, or have not been reduced to writing at all, so that they make little sense in an index of this kind. Also excluded is Rajasthani, which is in fact a bundle of closely related languages without a standardized orthography, and I am not sure about its official status.

Those languages currently available are written in bold face. You can expect more languages to be included here as time goes. Note that some languages are somewhat problematic to add, because they contain sounds/letters that cannot be mapped to the Devanagari alphabet. There are two supplementary letters in Bengali script (needed for Assamese language) that have no Devanagari counterpart, and Sinhala is even more problematic with its extra vowels and its prenasalisation feature; these letters have been conflated with standard Devanagari letters in sorting. For the same reason, Tibetan was difficult to add, as many ad hoc decisions had to be made how to align Tibetan script with Devanagari; readers are probably better served with the Tibetan Index.

The South East Asian Brahmi-derived scripts are perhaps impossible to add to this index, as they have evolved a long, long road from the common ancestor.

It should be noted that I obtained many of the names shown in this index from poorly legible, hand-written lists, and (as anyone with some knowledge of Indic scripts will confirm) it is quite difficult and error-prone to digitalize such scribblings, moreover one may easily fall victim to orthographic deficiencies of the indivuals who wrote them. Whenever possible, I checked on the Internet, but due to a scarcity of sources (especially bilingual ones), this was only partially possible for some languages (notably, Gujarati, Bengali and Malayalam). Spellings in Sinhala, Punjabi, Assamese, Nepali and Konkani should therefore taken with a grain of salt. Comments and corrections are of course welcome.

Some of the more exotic languages with null internet coverage were field researched, which poses significant problems with poor alphabetization rates and a general lack of writing tradition. Thus, spice names in Maithili, Dogri and Newari are perhaps to be seen as approximations only (the number of different spellings for a word can well equal the number of people asked); also, Tibetan proved hard work, although I was finally able to cross-check some of the names with written literatur on medicinal herbs.

Sanskrit presents a different problem, as the standard dictionary of Monier Williams contains an unreasonable amount of synonyms and polyvalencies (and invalid scientific plant names); neither ancient Indian writers nor modern linguists give much heed to botany, it seems. For that reason, Sanskrit will probably remain in the β state forever.