Writing แ, Not เ + เ

Thai แ is one vowel character: U+0E41 THAI CHARACTER SARA AE. Its two upright strokes may look like two copies of เ placed side by side, but that visual resemblance is not an encoding recipe. Standard Thai text stores แ as one code point, never as U+0E40 repeated twice.

This is not merely a programmer's preference. A look-alike string can break search, collation, cursor movement, spell-checking, accessibility tools, and text-to-speech caches while appearing acceptable in one font. The safe learner habit is simple: type แ from a Thai keyboard and preserve that exact character.

The form and its place in a syllable

Sara ae represents the vowel quality conventionally romanized ae in RTGS. Like เ, โ, ใ, and ไ, it is a spacing vowel character written before the onset consonant even though the consonant is pronounced first. It is not a combining mark above or below a base.

The display order in แกะ is แ + ก + ะ, but the spoken analysis is onset ก /k/ followed by short /ae/. The visual order is therefore not the sound order. In แม่, the stored order is แ + ม + ่: the preposed vowel comes first, the onset consonant next, and the tone mark follows its base in the encoded sequence.

One character, not a doubled vowel

The incorrect look-alike is a sequence of two U+0E40 SARA E characters. It is deliberately described by code point rather than printed here, because copying the malformed sequence would reproduce the problem. U+0E41 is not a compatibility decomposition, stylistic ligature, or font shortcut for that sequence: it is an independent Thai character.

This differs from English typography. English readers are accustomed to seeing two letters such as ee represent a vowel, so it is tempting to interpret the two strokes of แ as two units. Thai readers identify the whole glyph as one vowel sign. The same principle applies whether the text is (informal) chat, neutral prose, or (formal) legislation.

In ดูแล and แนะนำ, แ is one U+0E41 character wherever it occurs. By contrast, เมนู begins with the single SARA E character เ. The two signs are visually related but not interchangeable, and searching for a malformed doubled-U+0E40 sequence should never be treated as normal Thai spelling.

Syllable-level reading walkthroughs

Consider แพง “expensive”:

1. The visible sequence begins with แ, but พ is the onset.
2. พ is a low-class consonant.
3. แ supplies a long vowel and ง supplies a final sonorant, so the syllable is live.
4. There is no tone mark. The full tone rule yields a mid tone in standard Central Thai.
5. RTGS gives phaeng, omitting both tone and vowel length.

Now consider แปลก “strange”:

1. แ stands before the true cluster ปล.
2. ป is the first consonant of the onset and controls the consonant-class input.
3. Final ก is pronounced /k/, making the syllable dead.
4. The tone follows from the entire spelling pattern, not from แ alone.

Short and long forms are not interchangeable

The open-syllable contrast แ–ะ versus แ– normally distinguishes short and long vowels. Compare แพะ phae “goat” with แพ phae “raft.” RTGS writes both as phae because it does not mark vowel length or tone. The Thai spelling—and ultimately the word—carries information that the reading aid omits.

Before a pronounced final consonant, short /ae/ is commonly written with ไม้ไต่คู้ in the pattern แ–็–, as in แข็ง “hard; strong; cold as in ice.” The exact spelling is lexical; learners cannot recover it reliably from an RTGS form.

English spelling sometimes tolerates multiple letter sequences for the same sound. Thai likewise has historical spelling complexity, but that does not make its code points freely substitutable. The writing system preserves consonant classes, finals, vowel patterns, and marks that jointly predict a reading. A screenshot may conceal an encoding error; linguistic analysis will not make that error legitimate.

Search, OCR, and normalization

Unicode normalization does not turn two U+0E40 characters into U+0E41. A pipeline that “fixes” the malformed sequence only during rendering leaves the underlying data wrong. Validation should inspect code points at ingestion, after OCR, and before generating search indexes or audio.

OCR is especially vulnerable because the two vertical shapes invite segmentation as separate glyphs. The reverse error is also possible: adjacent independent เ characters across an unusual layout can be fused incorrectly. Repair must therefore consult the Thai word and syllable, not just replace shapes blindly.

Here แคบ contains one แ before ค. An OCR result that stores two copies of SARA E is not a variant spelling of “narrow”; it is corrupted text. A code-point validator can flag it, while a dictionary or language model can confirm the intended word.

The visual-order page develops the broader difference between storage, display, and sound order. The vowel-map places short and long ae in the full vowel system.

Historical and register notes

The shape of แ is historically related to doubling, but modern encoding and modern literacy treat it as a distinct vowel character. Historical origin does not license a doubled-U+0E40 spelling in current digital text. Use the single character in every register.

Manuscript transcription may record genuinely unusual historical glyph forms (academic/archival), and a diplomatic edition may preserve scribal variation. Such work should document its transcription policy explicitly. It should not silently substitute a malformed modern look-alike for U+0E41.

Common Mistakes

1. Typing the visible strokes as two characters

2. Pronouncing the preposed vowel before the onset

3. Assuming RTGS preserves vowel length

4. Treating แ as a tone instruction

5. Trusting appearance instead of code points

Key Takeaways

แ is U+0E41 and is always one character in modern Thai text. It is written before the onset but pronounced after it, and its short, long, closed-syllable, and tone behavior depends on the rest of the spelling. Never replace it with two U+0E40 characters, and do not trust a visually plausible rendering: validate the actual encoded text.