Nyaya Ratnamala of Parthasarathi Misra 4.1

Where they disagree regarding the relation of the sign and the signified,
both with respect to nature and means of valid cognition, there the determination is now stated.

First, regarding [its] intrinsic nature: followers of Kaṇāda assert that [the relation is] the cause of inference denoted by the word 'laiṅgika' (inferential sign) on account of causality, conjunction, inherence, inherence in a common object, and contradiction, saying: 'This is its cause, its effect, conjunct, inherent in a common object, and contradictory—this is what constitutes an inferential sign.' The Buddhists hold that invariable concomitance is the positive and negative concomitance of the probans with the probandum and its absence, rooted in essential identity and causation. Others hold it to be pervasion of the probans by the probandum. The Mīmāṃsakas hold it to be of the nature of a restriction. There is [likewise] disagreement regarding the means of knowledge [for grasping it]: the Buddhists state that essential identity and causation are the means of knowledge. Others say it is mental perception. Others say it is purely sensory perception. Some say it is even the initial perception. Still others consider the final perception, assisted by repeated observation, to be the means of knowledge for restriction. Others say it is postulation. The Teacher holds that it is repeated observations themselves, assisted by non-observation of counter-instances. Among these views, if one accepted only causality and so forth as adopted by the Vaiśeṣikas, there could be no relation of revealer and revealed between the rising of the Kṛttikās and [the rising of] Rohiṇī and the like, because causality and so forth are absent [between them]. Nor can all [relations] be included by taking the term 'conjunct' (saṃbandhin), because that refers merely to conjunction and inherence. And if it referred to everything, then causality and the rest would also be established from that alone, making their separate mention useless. Moreover, in this way, there would also be an inference of smoke from fire, since they are [related as] cause and effect; [or an inference of being a tree] from being a śiṃśapā tree, because they co-inhere in the same object. If [it is argued that] this unwanted consequence does not follow because it is not restricted, then restriction alone is the auxiliary factor for inference, not causality and so forth. Even on the Buddhist view, because the reason is an inferential mark only through obtaining positive concomitance of essential identity and effect alone, the cognition of Rohiṇī having recently risen would not be established from the rising of the Kṛttikās. And if it is said that the inference of color from taste is an inference of the material cause of color, which is a cooperative cause of the material cause of taste, from taste by virtue of the efficacy of action, and from that material cause of color an inference of the resulting color is made by virtue of the efficacy of action—this is not so, because [one] is not the effect [of the other]. Even though a cause together with its auxiliary causes is inferred from an effect, an effect cannot be inferred from a cause, even through the efficacy of action, because it is not an effect [of it] nor of the same essence. If a cause can also be an indicator, then, as stated, the unwanted consequence of inferring smoke from fire would follow. Furthermore, positive concomitance means following after; but fire, which is situated on the ground, does not follow after smoke that is rising upward; nor does the sun in the sky follow the shadow of a gnomon situated on the ground. By this, pervasion is also refuted, for: