What is the purpose of transliteration?

Question

On the Wylie Tibetan Transliteration page (original paper), it says:

Previous transcription schemes sought to split the difference with the result that they achieved neither goal perfectly. Wylie transliteration was designed to precisely transcribe Tibetan script as written, which led to its acceptance in academic and historical studies. It is not intended to represent the pronunciation of Tibetan words.

A google search for "what is the purpose of transliteration" reveals:

Transliteration is the process of transferring a word from the alphabet of one language to another. Transliteration helps people pronounce words and names in foreign languages. ... It changes the letters from the word's original alphabet to similar-sounding letters in a different one.

Those two answers are in conflict.

To me -- after looking at a few different transliterations -- it seems that transliterations exist because before there was computers it was hard to write and read the characters of the native language efficiently, so they used familiar symbols (latin letters) to "transliterate". But it seems that a good transliterator doesn't focus on pronunciation, but instead, does focus on perfect writing transcription or perfect copying of the letters so they can be recovered (because of the transliterators 1-to-1 mapping).

If this is the case -- that a good transliterator is one that doesn't focus on pronunciation but instead on writing transcription -- then why don't you just use a numeric code to represent every possible glyph in the native language? That is, why don't you just use some existing numeric code like Unicode! Which gives you the benefit that you can recover the original glyphs, but nowadays you can actually see the original writing on your laptop or phone because the fonts all now include it pretty much.

So my question is, is this correct? Were transliteration schemes created to solve the problem of manipulating a target language in easier to write and read (familiar) symbols, with the goal of perfect transcription of characters rather than sounds, because computers weren't present? Or was/were there some other reason(s)? Would it be advisable to use a transliteration scheme today, like the Wiley one? Or just go with Unicode? What is the benefit of one over the other? Are transliterations obsolete because of Unicode?

Of Sanskrit:

The International Alphabet of Sanskrit Transliteration (IAST) is a subset of the ISO 15919 standard, used for the transliteration of Sanskrit, Prakrit and Pāḷi into Roman script with diacritics. IAST is a widely used standard. It uses diacritics to disambiguate phonetically similar but not identical Sanskrit glyphs. For example, dental and retroflex consonants are disambiguated with an underdot: dental द=d and retroflex ड=ḍ. An important feature of IAST is that it is losslessly reversible,[citation needed] i.e., IAST transliteration may be converted to Devanāgarī or other South Asian scripts without ambiguity and with correct Devanāgarī spelling. Many Unicode fonts fully support IAST display and printing.

As a bonus question, what popular languages don't have a losslessly reversible transliteration scheme? It seems that Korean does not. Chinese seems too complicated I don't know. (Japanese, Mongolian, Russian, Telugu, Tamil, Thai, Cree, Georgian, Armenian, Amharic, Hebrew, Arabic, Persian, Avestan, Ancient Egyptian, Aramaic, Coptic, Gothic, Old English)

I ask because of this:

• Written form: exact replication of letters (unicode is best for this)
• Phonetic form: exact replication of sounds (IPA is best for this)

Which is why to get the best of both worlds, you can't have 1 language, you must have two.

Answer 1

The words "transliteration" and "transcription" are often used interchangeably. If you want to draw a distinction between them, the distinction is usually that "transliteration" tries to encode the original grapheme sequence as losslessly and efficiently as possible, while "transcription" tries to encode the original phoneme sequence as losslessly and efficiently as possible. Sometimes people call the latter "bound transcription" for clarity.

In the age of Unicode, why do we need transliteration? Well, most people can't read raw Unicode codepoints. If you want to find a particular poem in a library's collection, is it easier for you to identify 631 62b 627 621 20 627 644 623 646 62F 644 633 by 627 644 631 646 62F 64A, or رثاء الأندلس by الرندي, or the Rithā' ul-'Āndalus by Al-Rundī? Most English-speaking librarians would find the latter significantly easier to catalog, sort, file, reshelve, etc.

In the pre-Unicode days, it was also often impossible to typeset or print various exotic writing systems. Nobody was really putting in the time and effort to make movable type for Hittite cuneiform, and many Indo-Europeanists weren't fluent at reading it anyway. So instead of writing the word 𒈾𒄴𒋼𒂊𒉌 in their papers, they would use the transliteration na-aḫ-te-e-ni (representing each sign unambiguously), or the bound transcription naḫtēni (representing the phoneme sequence rather than the signs). Even nowadays, if you've never had to deal with cuneiform before, it'll be much, much easier to learn to read naḫtēni than 𒈾𒄴𒋼𒂊𒉌.

(Side note: even now, Unicode hasn't solved all the problems! I can't actually read the word 𒈾𒄴𒋼𒂊𒉌 in my own answer—because the default font on StackExchange has classical Sumerian cuneiform glyphs, which look very different from the modified Old Assyrian glyphs used for Hittite! Without some way to change the font, I have to copy-paste into a text editor and switch to a Hittite font in order to get something I can read:

Naḫtēni and na-aḫ-te-e-ni meanwhile are as readable as ever.)

Answer 2

One can actually read a transliterated text. Hardly possible with a sequence of Unicode numeric codes. If you want the most exact replication of the text, then even Unicode might not be "best" (your wording). You might need a photograph or an accurate drawing.

Often you do not need the details of the graphics, you need to reproduce the sequence of letters in a way that is easier to work with, that you and you readers can read easily. Slavists will customarily transliterate Glagolitic texts to Cyrillic or Latin so that you can fluently read the text and work with the meaning of those words as well.

In certain conventions one might loose the information like the rare presence of a spidery CH or the multiple Is present in glagolitic (vs. eg. Cyrillic single и and single х) and then the process is not strictly reversible and may be instead called transcription. It allows much easier work with those texts. Also getting rid of those redundancies will actually make searching of words contaning those words much easier. Or you can choose strict transliteration to preserve the distinction with the use of і,и and so on.

The pronunciation problem is secondary. With a faithful transliteration the pronunciation rules for the sequences of the original script graphemes will remain the same for the transliterated graphemes.

Answer 3

Transliteration is useful because linguistic research is meant to be accessible to linguists generally and not just people with some prior experience with the language in question. Transliteration allows people to read and understand articles without needing to become proficient in a new script or language. We shouldn't assume that everyone will be familiar with a language or script before they read our research; this is especially the case with typology where you might be searching for papers by keywords and reading research on a dozen new languages every day.

To give one concrete example, in the Hebrew alphabet there are three pairs of letters which can be easy to mistake for each other: a: ב and כ, b: ד and ר, and c: ה and ח. In some fonts, particular classical serif ones, they will look much more similar than they do here on this web site. Even if you know the language you will probably still occasionally confuse them, especially if you're reading small text or a poor photocopy. (Even the Hebrew Bible has mistakes from these characters being confused!) If you misread these characters you could completely misunderstand an article about Hebrew phonology, or considering that two of the three Hebrew prepositions are ב and כ, you could be entirely bewildered if you didn't realise they were completely different characters!

Now readers do have a responsibility to do their own work to understand what they're reading. But providing transliterations is a relatively small thing to do which could have a massive impact on your research's accessibility, and thereby on its use and impact.