Another reason for the constitution of the interjection as a separate grammatical class is their specific mode of signifying, which prefigures the dominating trend of middle Ages speculative grammar. In a late document, of the XII^th – XIII^th century (Commentum in Artem Donati), primary and secondary interjections are regrouped by a single definition based on the basis of their common substance, which for the speculative grammarians is a distinct mode of signifying.

            The tradition of Latin grammarians perpetuates the idea of an unarticulated or unachieved mode of signifying (voce incondita, informata) that would be specific to the interjection. This unachieved articulation would reflect, according to these writers and to the philosophical tradition, the disruptive effect that feelings may have on intellect, on reason, on the faculty of speech:

            The observation of such an "unachieved" linguistic form has determined early grammarians to approach the elements labelled as interjection as a unique, and anomalous linguistic phenomenon. It is the reason why Greek grammarians banished from language, under the label alogoi, some of the items inventoried by later or contemporary linguists as interjections. Priscian, in the IVth century, notices the fact that interjections make use of unarticulated (illiterati) sounds, and adduces in support to this statement the examples of the onomatopoeia imitating laughter (hahaha) or the interjection phy, euhoe, au [7].

            Priscian, as well as numerous other authors, pointed out the accentual instability characteristic to interjections and underlined its anomalous character as far as stress patterns are regarded, which is connected to their unachieved, unformed nature.

            Or

            Cledonius links the accentual instability of interjection to their formal variability:

            Such and similar observations, together with remarks related to the syllabic structure, preponderantly monosyllabic, or to the vowel length of interjections, can be found in a series of authors among which: Donatus, Cledonius, Maximus Victorinus, Cassiodor, Commentum Einsidlense.

            An original aspect related to the anomalous character of interjection is discussed in De ultimis syllabis (erroneously attribute to Probus). The author provides a possible distribution scheme for interjections in the various literary genres. Accordingly, interjections are found to be more frequent in drama and short lyrics (carmina). On a metric basis, and not for semantic reasons, the anonymous author considers that a trisyllabic interjection, consisting of long vowels, could only be acceptable in the comic genre:

            Other interjections, such as va(h), (e)heu, papae, nefas, due to their metrical structure and their expressive function are compatible with the lyric, or even oratorical genres:

            The author of Regula Aurelii Augustini, noticing the "unachieved" form of the interjections, adopts a radical position and proposes the exclusion of such anomalous items from the language. He inaugurates the long debate among grammarians and linguists concerning the acceptability of this new created word class as linguistic or not.

            Diomedes underlines the unique nature of the mode of signifying specific to interjections, voce incondita. He discusses another controversial aspect that contributed to the "anomalous" nature of interjection, namely the double resources of their meaning: a codified, conventionalized meaning (ex consuetudine) and a contextually actualized meaning (ex sequentibus verbis). Another major debate is thus prefigured, related to the semantic content of interjection. There are authors sustaining the possibility of reducing to a unique semantic nucleus all the occurrences of a given interjections, and on the other side, there are authors that deny interjections a stable meaning, and on this basis propose their exclusion from language. The idea that interjections need to contextually specify their full meaning coexists paradoxically with that of its contribution to disambiguate speaker's attitude towards the informative utterances accompanying the interjections, as in the examples and the discussion provided by Sergius and Pompeius. These two authors underline the important function of interjections at the communication level, for expressing affects and attitudes in a way for which the referential discuss does not possess sufficient resources.

            Another connected reason for treating interjections as anomalous units is their apparent syntactic independence. The disambiguating role of interjections that function as attitudinal indicators or comments on adjacent informative utterances suggests the special nature of the relationships between the two types of linguistic units. Given the absence of explicit marks, at the surface level, traditional grammar posited the impossibility for interjections to enter in syntactic relationships with adjacent units. Priscian (Instit., lib. XV, § 40) comments on this, already in the IVth century. He links the syntactic isolation of interjection to its affective motivation: