First of all: don’t get your knickers in a twist!
Every phrase of every sentence of a translation requires judgment, care, and reflection, and there is no one “right” way to do things. Translation of Pali verbs into English requires sensitivity to context, as there is never a one-to-one mapping of forms from one language to another. Indeed, the very idea of a grammatical form is merely an abstraction, a concept invented by grammarians to help understand language. Abstractions developed in one language only ever loosely apply even to that language and comparing the sets of abstractions and conventions (AKA the “rules of grammar”) between one language and another is never more than a broad approximation.
Thus when choosing to translate the first person indicative verb, or the present active participle, a translator applies their best judgement to how to express it in English. Making one choice rather than another by no means constitutes a “major error”. And to depict it as that is not remotely accurate or helpful.
Now to explain what is happening.
The Pali text has a series of similar phrases, one for each of the four postures. Normally in sets like this the different phrases are presented with a high degree of syntactical consistency. But this passage is a little unusual in that the syntax varies in the four cases.
In the final three cases, i.e. standing, sitting, and lying down, the sentence uses a doubled phrase, where the past participle is echoed with the same form, the past participle. Literally it is something like “One who has stood knows ‘I am one who has stood’.” The past participle is used in such cases in a perfect sense, to indicate that the action has been completed. Each of these postures is a static one, and one is in that posture. Here, the verb “I am” (amhi) is applied to the past participle, as the past participle verb form does not specify person.
With walking, however, that phrasing cannot be applied. Why? Because the past participle means “gone”. Obviously this doesn’t work: “One who has gone knows ‘I am one who has gone’.” So in the first part of the phrase, the present participle is used (gacchanto “going”), and it is echoed with the present indicative verb, “I go”. Lit: “The going one knows ‘I go’.”
In this case, the person “I am” need not be specified by amhi, as it is already coded in the first person verb form.
Thus the difference in syntax between the first phrase and the next three is entirely due to the specific nature of Pali syntax. No translator would waste their time trying to replicate that, as in any case, the sentence must be expressed differently to be idiomatic English. Note, however, that in my translation, the different syntaxes are, in fact, echoed so very slightly in my phrasing. You’re welcome!
when a mendicant is walking they know: ‘I am walking.’ When standing they know: ‘I am standing.’ When sitting they know: ‘I am sitting.’ And when lying down they know: ‘I am lying down.’