In fact, the emphasis on the subject as knower might be interpreted as just instrumental to the ontological foundation of the ātman. That this is not the case is at least suggested by the history of the Mīmāṃsā reflection on this subject, which originates from Vedic hermeneutic concerns and not from ontological ones. Historically, it is more probably the case that ontological arguments have been added in order to firmly ground the subject's phenomenological role in epistemology and ritual hermeneutics.
How comes, then, that the ontological approach is added to the phenomenological one? Hypothetically, one might propose that Śālikanātha adds to the standard Prābhākara argumentation, based on the Veda and, therefore, on the phenomenology of the subject and on its role in epistemology, a further line of argumentation, based on ontology and inspired by the arguments of its opposers, mainly Naiyāyika authors. Consider the following sequence of arguments (with no interruption in between) and note the shift to the usage of ātman in the ontological section:
(Vedic-based argument:) In regard to someone who [performs] no action, agency and experience would be [merely] hypothetical. [But] an agent and an experiencer, presupposed by the [Vedic] sentence about the one having the sacrifice's weapons [which refers to the ātman as the sacrificer and as the one who will enjoy heavenly bliss],1 must be determined, hence the mention of recognition has been put forth.
(Nyāya-influenced argument:) And when this self has been recognised, then the self [endowed with] consciousness is its characteristic. Nor does it arise at a certain point from it (self) alone. For an effect whose [single] cause is always proximate would not arise [only] from time to time. […] Therefore this self, which is the inherent cause [of consciousness], requires a [further] cause, one that is inherent to it.2
1 “This sacrificer who has the sacrifice's weapons (the sacrifice's substances and tools), will go to the heavenly world quickly” (sa eṣa yajñāyudhī yajamāno 'ñjasā svargaṃ lokaṃ yāti). The sentence is discussed by Śabara, ŚBh ad MS 1.1.5. I could not locate a Vedic source for it.
2 According to the Nyāya school, there are three sorts of causes: 1. the samavāyikāraṇa 'inherent cause', such as the two halves of a pot, which inhere in the effect (the pot); 2. the asamavāyikāraṇa 'cause which does not inhere [in the effect, but rather inheres in the inherent cause], such as the colour of the two halves, which inheres in them, but not directly in the pot; 3. the nimittakāraṇa 'necessary condition', such as the potter.
Why does Śālikanātha feel the need to add ontological arguments? (My tentative answer will follow shortly, but I would be glad to read your opinion about it).
On more insights about the subject (this time in the Mānavadharmaśāstra) by Daniele Cuneo, see here.