Writing without an alphabet  

Pali is a phonetic language with no written alphabet of its own. Students of the language have therefore relied on their own native alphabets to read and write Pali, ever since the 1^st century BCE, when Sri Lankan scribes first recorded the Tipitaka in the Sinhala alphabet. But when Europeans began to take an interest in South Asian languages in the 19^th century, they quickly discovered that their own roman alphabet was no match for the wide range of phonemes (sounds) present in South Asian languages. European scholars thus began representing the more problematic Pali phonemes by augmenting the roman alphabet with a system of letter-pairs and diacritics, including the macron (horizontal bar), dot-over, dot-under, and tilde:

Until well into the mid-20^th century, Pali typefaces using these characters were used almost exclusively by specialty book publishers; a scholar's day-to-day duties of transcribing, translating, and editing had to be laboriously carried out with typewriter, pen, and a steady hand with which to apply the diacritics. Unfortunately, the first personal computers failed to address the typographic challenge of diacritics, as they were designed around a very limited character set (ASCII) that was only barely able to accommodate the upper- and lower-case roman letters, ten digits, and a modest sprinkling of punctuation marks. The extended-ASCII set, which soon followed, offered a suite of additional special symbols, including many required for northern- and eastern- European alphabets. But still no macrons or dot-unders. In the absence of a universally accepted computer representation of non-ASCII characters, students of non-European languages were left to invent their own stopgap methods. These range from giving ordinary punctuation marks double-duty as stand-ins for diacritics, to designing special diacritic fonts (all of which are incompatible with each other), and everything in between.

Evaluating the methods  

A good written phonetic representation of Pali — indeed, of any language — using one's native alphabet as a starting point should aspire to each of the following ideals:

1. It should be readable by a wide audience. It should introduce a minimum of special characters that are not already present in the alphabet. It is better to modify an existing letter with a small diacritic than to introduce an entirely new character that may look like an alien squiggle to the uninitiated. A newcomer to Pali, upon seeing a t with a dot-under, should be able to guess immediately that the letter stands for some variant of a t sound.
2. It should be phonetically precise. The written text should precisely and accurately capture the phonetic content. Each phoneme (sound) should be unambiguously represented by a unique letter or combination of letters.
3. It should be easy to type. Writing Pali should not be a cumbersome exercise in keyboard gymnastics. Typing an a-macron should not call for a long series of keystrokes (e.g., Alt-Ctl-Shift-Esc-a).
4. It should be portable. If you hand me a book — or send me a text file by e-mail — it should appear to me exactly as it did to you. I should be able to sound out the text phonetically exactly as you intended.

No single method simultaneously realizes all of these goals; no single method is "best" Your method of choice therefore depends both on your particular needs (e.g., Do you demand phonetic precision? Are you printing a book or dashing off a quick e-mail?) and on the typing, printing, and computing resources at your disposal (e.g., Do you have a Pali font? Does your text-editing app support Unicode?).

In what follows I've singled out some of the more common strategies that Pali students have used in recent decades, running the gamut from ignoring diacritics altogether to using Unicode fonts. I evaluate the success of each strategy in achieving the above-mentioned goals, to help you decide which method best suits your needs.

Method 1. Ignore the diacritics  

This is certainly the simplest method. But the cost of that simplicity is heavy: the irretrievable loss of crucial pronunciation details. This is the method often used here at Access to Insight. (I should add that I do make use of the palatal nasal ñ because it is so easy to implement using HTML and because it is contained in the extended-ASCII character set found on practically everyone's computer nowadays.)

Method 2. Use capital letters  

Capitalized letters represent letters with an accompanying diacritic. The method is simple, but it has ambiguities: how, for example, would you distinguish between the palatal and guttural n (n with a dot-under, and n with a dot-over)?

Method 3. The Velthuis scheme: double the vowels, punctuate the consonants  

In the Velthuis scheme two basic rules are observed:

1. Long vowels (those usually typeset with a macron (bar) above them) are doubled: aa ii uu
2. For consonants, the diacritic mark precedes the letter it affects. Thus, the retroflex (cerebral) consonants (usually typeset with a dot underneath) are: .t .th .d .dh .n .l. The pure nasal (niggahiita) m, also typeset with a dot underneath, is .m. The guttural nasal (n with a dot above) is represented as "n . The palatal nasal (n with a tilde) is ~n.

Of the plain-ASCII methods, this one is the most precise, as it carefully preserves the significance of each special character. To the uninitiated, however, the sight of all those doubled vowels and misplaced periods is utterly bewildering, perhaps leaving them to wonder if someone's keyboard is broken.

Method 4. Use a little HTML  

HTML has access to the extended ASCII character set, which includes many accented non-English European vowels (umlaut, circumflex, etc.), some of which can serve as reasonable stand-ins for the long Pali vowels (ä ï ü; à ì ù; or â î û etc.). The palatal n is straightforward: ñ. Whatever type of accent you adopt, use it consistently.

Method 5. Mixed Velthuis/HTML  

This method attempts to clear up the stuttering of Method 3's doubled vowels, by using a little HTML (Method 4).

Method 6. Special Pali fonts  

For high-quality print projects, nothing beats a well-designed Pali font. For an extensive review of available Pali and Sanskrit fonts, see Transliteration and Devanagari Fonts for Sanskrit, by Ulrich Stiehl. The Association for Insight Meditation's Pali Font Resources page offers several ANSI and Unicode fonts suitable for working with Pali.

Example (in "Normyn" font):
Readability:	Excellent.
Phonetics:	Excellent.
Ease of use:	Variable — it depends on the keyboard mappings used by a particular font.
Portability:	Poor. These fonts don't all share the same coding standards; they are not interchangeable. If I send you a text document that I formatted with font X, and you display it with font Y, the Pali characters may not show up properly.
Overall:	Excellent — but only for documents that are to be shared in print (hard copy) form or as PDF files or GIF images.
Uses:	Printing. Not suitable for e-mail or the web, except when embedded in PDF files or GIF images.

Method 7. Unicode and Unicode fonts  

Unicode has emerged in recent years as the international standard for representing characters from most of the world's alphabets. All the special characters we need for Pali transliteration may be found in Unicode's Latin Extended-A, and Latin Extended Additional code charts. They can therefore be easily generated using HTML, provided that your web browser uses a Unicode-savvy font.

There are many Unicode fonts available that contain the characters needed for Pali.

The following table lists the HTML Unicode entities required to generate each of the special Pali characters. If your web browser supports Unicode, the characters appearing in the last column of the table should resemble those appearing the shaded column. If they do not match, then you may have to upgrade your web browser, install Unicode fonts on your computer, or both. For details about configuring your computer and browser to use Unicode, see the Unicode website.

Pali letter		Velthuis	HTML	Rendered on your browser as [3]
A macron		AA	Ā	Ā
		aa	ā	ā
I macron		II	Ī	Ī
		ii	ī	ī
U macron		UU	Ū	Ū
		uu	ū	ū
N dot-over		"N	Ṅ	Ṅ
		"n	ṅ	ṅ
M dot-under		.M	Ṃ	Ṃ
		.m	ṃ	ṃ
N tilde		~N	Ñ	Ñ
		~n	ñ	ñ
T dot-under		.T	Ṭ	Ṭ
		.t	ṭ	ṭ
D dot-under		.D	Ḍ	Ḍ
		.d	ḍ	ḍ
N dot-under		.N	Ṇ	Ṇ
		.n	ṇ	ṇ
L dot-under		.L	Ḷ	Ḷ
		.l	ḷ	ḷ