Why breath meditation has four, six, or eight stages

What does a meditation manual count?

A list of eight can appear to contain twice the practice of a list of four. When both begin with counting and following the breath, it is tempting to imagine a simple history: an earlier method gradually acquired more stages. Yet the number of headings cannot establish that history. We first need to ask what each heading names, and what the explanation beneath it contains.

The Chinese Vimuttimagga, or Path of Freedom, introduces four ways of developing mindfulness of breathing: counting, following, fixing, and observing. The received Chinese Ānāpāna text, T602, uses six terms: counting, following, stopping, observing, returning, and purifying. Buddhaghosa’s Visuddhimagga gives eight. These are useful working translations, but similar English words do not establish identical meanings across the manuals.^[1][2][3]

The instruction that a list can hide

Breath-contact provides a revealing test. The Visuddhimagga names touching, phusanā, as a separate member of its list. The Vimuttimagga’s four headings do not. A comparison limited to those headings might therefore suggest that one manual added a practice missing from the other.

But the Chinese Vimuttimagga describes establishing attention where the breath touches within its explanation of fixing. Contact is already there, even though it has no heading of its own. The Visuddhimagga adds another complication: it explains that touching and fixing do not require separate acts of attention in the same way as counting and following.^[1][3]

The difference is partly about how an instruction is organized. An operation can be present inside another explanation, while a separately numbered item need not describe a separately performed exercise. Counting the names is a starting point for comparison, not its conclusion.

Where the comparison leads

The larger study asks how the manuals organize attention and how their lists connect practical instructions with insight, attainment, and review. It also examines T602’s instruction concerning the nose-tip. None of these correspondences, by itself, establishes which text borrowed from which, or places all three on a single historical ladder.^[2][3]

This opening introduces a question developed from the appendices to The Scripture on Ānāpāna and our work on the Vimuttimagga. The full paper sets out the Chinese and Pāli passages, a comparative table, and the limits of the proposed correspondences.