Do Not Insert Spaces between Every Thai Word

Thai normally does not print a blank between every lexical word. Words inside a phrase run together, while a visible space can separate larger units: a sentence, a substantial clause or sense group, items in a list, or fields in mixed-script and numeric text. The exact placement is partly editorial; short tightly connected clauses often remain continuous.

The English transfer error is a category error. An English space is usually a word boundary. A Thai space often behaves more like punctuation or layout. Copying every RTGS space upward into Thai therefore fragments compounds, classifier phrases, particles, and predicates. Removing every Thai space is also wrong, because larger information units then collide.

For the positive system, see direction and Thai spacing and Thai punctuation. This page focuses on diagnosing imported English boundaries.

The unit to preserve is the Thai construction

In หนังสือสามเล่มนี้, the noun หนังสือ, numeral สาม, classifier เล่ม, and demonstrative นี้ are separate grammatical pieces, but ordinary Thai prints the phrase continuously. In กินข้าวแล้วค่ะ, the object, aspect particle, and polite particle also remain attached in normal prose.

RTGS on this site deliberately inserts learning spaces: nangsue sam lem ni. Those spaces help expose words but are not a template for Thai orthography.

Common Mistakes

1. Spacing every dictionary word

The rejected lines are still decipherable and may appear in intentionally segmented beginner materials. They are not normal Thai prose. That special teaching context does not license copying the segmentation into a message, form, or sign.

2. Splitting compounds and classifier phrases

The classifier does not receive English-style spaces around it. The phrase remains readable because Thai grammar predicts the sequence noun + numeral + classifier + demonstrative. In a table or vocabulary exercise, an editor may separate columns, but that is layout, not sentence orthography.

3. Detaching particles and tightly bound predicate material

The Royal Society convention normally gives ๆ visible space around it in running text: ช้า ๆ, แปลก ๆ, เด็ก ๆ. This is not evidence that every word should be spaced. It is a convention attached to a particular symbol.

4. Deleting spaces that organize complete thoughts

The rejected versions are not automatically ungrammatical. Thai writers have discretion, and short connected clauses may run together. Here the corrections express the intended rhetorical break more clearly. Diagnose information grouping rather than counting words.

Mixed text and names are separate editorial cases

Thai spacing also organizes fields that are not ordinary internal word boundaries. A form may show เวลา 10:00 น., a model label may combine Thai with Latin letters, and a full personal name normally has a space between given name and surname. These conventions do not turn the surrounding Thai phrase into English-style tokenization.

A spacing error causes a referential detour

A learner copies word boundaries from a gloss while asking a colleague to buy กล้วยแขก, the established name for banana fritters. Detached แขก can instead invite the reader to search for its independent meanings, including “guest.”

The problem is not that a blank always changes Thai meaning mechanically. It is that the learner displayed an internal English-style boundary and made the lexical grouping harder to recover. Context and the following noun ขนมทอด repair the intended referent.

Editing checklist

1. Ignore RTGS spaces when editing the Thai line. They represent learning units, not mandatory blanks.
2. Keep compounds together. Verify lexicalized units such as โรงพยาบาล and กล้วยแขก in a reliable dictionary.
3. Keep grammatical clusters together. Check noun–number–classifier–demonstrative phrases, verb–object units, aspect markers, and final particles.
4. Mark genuine sense breaks. Consider a space between a substantial subordinate clause and main clause, or between contrasted complete thoughts.
5. Apply symbol conventions. Check spacing around ๆ, abbreviations, numbers, names, parentheses, and list items rather than extrapolating one universal rule.
6. Read aloud by phrase. If every visible blank creates a stop, does the information structure still sound natural?
7. Check for invisible separators. Ordinary learner-facing Thai should not contain hidden word-boundary characters.

Key Takeaways

Thai spaces organize discourse and layout more often than lexical words. Normal phrases stay visually continuous, including compounds, classifier constructions, predicate material, and final particles. Preserve visible spaces where they clarify larger sense groups or follow a specific symbol convention, but never manufacture them by copying an English or RTGS word count.

Sources

• The Royal Society of Thailand, หลักเกณฑ์การเว้นวรรค, gives official spacing principles and examples, including sentence, clause, name, numeral, and repetition-sign conventions.
• Winskel and colleagues, “Effects of Interword Spacing and Morphological Complexity in Reading Thai”, documents normal unspaced Thai orthography and how added segmentation interacts with compound processing.
• Winskel, Radach, and Luksaneeyanawin, “Eye Movements when Reading Spaced and Unspaced Thai and English”, compares word-boundary processing in Thai and English and notes the punctuation-like role of Thai spaces.