I use two different transliteration schemes, depending on my mood. These I describe here. For names of my generation and the later ones, I have sometimes (read rarely) just followed the spelling used by people when they write their name in the Roman script: these are often extremely idiosyncratic, often attempting to give a very bad approximation to the true sound of the name and are directed towards a native speaker of Indian English (the vowel sounds are, thus, closer to those in the standard British dialect than in the American one); unless of course the parents live outside India in which case the spelling is tailored to the appropriate dialect. Some of these names are obviously names of English origin, and describing their bengalee spelling would be quite pointless. Their pronounciation, however, is often very Bengalee; so beware.
Bengalee uses an alphabetic script in which each Consonant(s)-Vowel syllable is often treated as an indivisible unit. Trailing consonants and leading vowels are written separately. The alphabet used is based on the Sanskrt alphabet: though the present sound values are very different.
The alphabet will be transcribed here as follows (it will be obvious from the context which of the two schemes below is being used).:
a A i I u U R e ai o au aM aH
These vowel sounds can be described in translation of terms that the ancient Sanskrit grammarians used: a - short closed guttural, A - long open guttural, i - short front palatal, I - long front palatal, u - short front labial, U - long front labial, r - short front alveolar (or retroflex), e - long back palatal, ai - a+i dipthong, o - long back labial, au - a+u dipthong, aM - a+closed guttural nasal, aH - a+rough breathing. Sanskrt had additional vowels RR (long front labial), L and LL (both dental like l) conventionally placed after R; out of these only the L is still counted in the Bengalee alphabet (or was counted when I was a kid: I think they have dropped it now), and none are used. The vowels i,u,e, and o had roughly the values in the English words hit, put, late and wrote, though in modern Bengali all are short (i.e. try let instead of late etc.). The long vowels (except aa) are the same as the preceding short vowels in modern Bengali. Furthermore, in modern Bengali, the vowel e is sometimes pronounced as the a in English cat, but this is rare for the e that appears in names. The vowel a has suffered a lot of change in modern Bengalee. In the initial syllable, it is usually pronounced as the English vowel in cot, in final positions (or final position of a combined word inside a compound) it is often omitted unless in a syllable with multiple consonants. In other positions or when following multiple consonants, it often becomes indistinguishable with o. A is pronounced roughly like the vowel in English car. In Bengalee, the vowel R is pronounced like the consonant r followed by the vowel i, often with an extra puff of breath: somewhat like rhi. Bengalee does not make a short/long difference for e and o that is found in some of the South Indian languages.
The consonants are (again in two different transcriptions):
The first five rows are ‘guttural’ (English k-hard g), ‘palatal’ (English ch-hard j), ‘alveolar’ (somewhat harder than the English t-d), ‘dental’ (more like the french t-like english th in the) and ‘labial’ (english p-b). The columns in these rows are ‘unvoiced unaspirated’, ‘unvoiced aspirated’, ‘voiced unaspirated’, ‘voiced aspirated’ and ‘voiced nasal unaspirated’. (Thus, for example, gh would sound like the english hard g with an extra aspiration and G is like the ng in sing). For t not folowed by a vowel, a distinct letter () is used, no distinction is made here. Except as a part of the last syllable of a word or a grammatical unit (when a slightly different letter is used in the Bengalee script, Y in the second transliteration scheme), y (palatal liquid) in Bengalee is pronounced as j if not combined with another consonant (or combined with r): in combination, it only serves to double the previous consonant. The combination yA when combined with a preceding consonant group is pronounced as the vowel in English cat. v (labial liquid) is pronounced indistinguishably from b in Bengalee: except in combination with preceding consonant (unless it is an r), it doubles that. The consonant r (alveolar, retroflex liquid slightly harder than English r) is either preceded by a vowel, or followed by one: neither of these occurs for the vowel R; same is the relation between l (dental liquid, just like English l) and L; In Bengalee, S (alveolar/retroflex sibilant: somewhat of a cross between kh and sh), s (dental sibilant like English s), and sh (palatal sibilant like English sh) are often pronounced indistingushably (like sh), as are n and N (like n). The combination, jJ () is pronounced the same way as gy would have been in the same position. The combination kS () is pronounced as kkh, and is traditionally considered a distinct letter. The consonant following consonant r is often written doubled, it is pronounced the same way whether doubled or not (a slight doubling can be heard). hm () often becomes mh in rapid speech. Otherwise, m following a group of non-nasal non-liquid non-sibilant consonants serves to double the value of those consonants, sometimes with a tinge of nasalization of the following vowel. Though G, and rarely J, do occur uncombined and have their true value in those contexts (though J often sounds like a nasalized y in common speech), most often they are combined with another consonant and is pronounced very similar to ng and nj respectively. The same holds true for M followed by a consonant in the first five rows: it gets replaced by the corresponding nasal. M followed by consonants from the last two rows, or alone, is pronounced as ng. Bengalee does not use the retroflex varieties of l that are found in some South Indian languages.
In addition, there are rolled versions of D and Dh which I write as D. () and Dh. (), or d__ and dh__. There is also a nasalization possible for vowels (e.g. ): written here as a ~ following the vowels.
I use a - (or .) to separate units that behave as separate words for pronounciation purposes when they are not written separately; though I have not done this consistently. Words often run into one another without separation in the manuscript: I do not indicate such mergers.
Incidentally, even if you can see the symbols written in bengali script, do not think you can learn to read or write bengali from this page. In bengali writing consonant clusters and the following vowel combine, so that one has to learn a lot of these symbols separately. Maybe someday I shall write a primer, but it is not ready yet.