I am often busy correcting my students’ essays and I often notice, when they decide to transcribe Sanskrit texts, that they have no idea about where to introduce white spaces. Hence, I thought of explaining it here, in case other students may need help with that.

1. Be clear about what you are doing: You are transcribing a text from one writing system to another. It does not make sense to retain in the new writing systems peculiarities which only make sense in the original one.
2. Therefore, you don’t need to reproduce in Roman alphabet the white spaces of the Devanagari (or of any other South Asian alphabet). For instance, mām eva, not māmeva, although the Devanagari would have मामेव.
3. Just use spaces between words.
4. What is a word? Use the Grammarians’ definition “whatever has a verbal or nominal ending” and add that the ending has to be visible (no lopa admitted).
5. Accordingly, you don’t need a white space within the elements of a compound. For instance, kṛṣṇacaraṇāravindayugalaṃ vande and not kṛṣṇa caraṇa aravinda yugalaṃ vande.
6. What about en-dashes, you might ask? You can use among elements of compounds, in order to make the life of your audience easier, especially in the case of longer compounds. Just be consistent and remember that the actual ending of the members of a compound are not visible. For instance, śrī anavadhikānantaprītisahitā can be transcribed as anavadhika-ananta-prīti-sahitā, but NOT as anavadhikā-ananta-prīti-sahitā.
7. Some authors use conventions to show that a sandhi has taken place. For instance, anavadhikânanta, meaning that there has been a sandhi of a+a.

Good luck with your essays and transcriptions!