I asked some time ago our Indo-Europeanist F. López-Menchero about some common (Classical) substantive endings we had to deal with in our Grammar. Here is his provisional answer before we make some corresponding corrections in our work.
1) About the Indo-European adjectives in -ak and -ik (viz. Lat. sagax, uiuax, felix; O.Ir. delech, bolach; Goth. wulthags, mahteigs; etc.), they have been studied quite well, there is a general consesus as to its original Indo-European form, and our Grammar doesn't show any 'independent' view: all are reconstructed as Eur. -ak or -ik.
2) About the -tāt ending, which was dealt with a year ago (version 2.11 of the Grammar), we decided to restrict some Latin extensions to its minimum extent. We considered the Spanish example "mur-ciélago", 'bat', which comes from "mur-ciego", 'blind mouse', which evolved from "-ciego" to a derivative "-ciégalo" and then to "-ciélago". It is obvious that some common modern words introduced by the Latin influence are dialectal forms that - we think - shouldn't be included in a common Indo-European language, unless there is a need for it. We deemed this -tāt ending a case for "Indo-Europeanization" of the classical words, like "universality", Lat. universitas, universitatis, or "manhood", Lat. uirtus, uirtutis, which we reproduced with a simple ending Eur. -tā, or -tus.
However, as López-Menchero points out, the forms in -tāt were productive not only in Greek and Latin, but also extensively in Old Indian; in fact, the most representative word is the one which referred to the whole, Gk. oloths, O.Ind. sarvā́tat(i)- (Pokorny 1819). It was possibly a group of abstract nouns related with the well-being or the situation, status or respect of a person or people that was generalized later in Greek and Latin to disparate terms referring to obscenity or punctuality.
Its declension swinged between the stems in consonant and those in -i like the diktis-type, as can be seen in Old Indian and some plural genitives as parisyllabic like Lat. ciuitatium, that vacillates with ciuitatum.
They are, indeed, supposed to be lengthenings from nouns in -tā (and some in -tūt like Lat. iuuentūs opposed to iuuenta - there existed maybe at some time iuuentus Gen. -ūs), just as Latin had pairs of the type augmen/augmentum, documen/documentum, frumen/frumentum, momen/momentum, tegmen/tegmentum, tormen/tormentum, segmen/segmentum.
One more proof of the relationship between -tāt and -tā is the formation of adjectives like "universitary" or "trinitary", where the second t doesn't appear.
3) Another curious stem is that of nouns in -tor/-ter. It was possible applied only to simple roots, like uisor, captor, dothr/dotwr. For compounds its homologous formation was a dental suffix -t, as in Latin sacerdos, antistes or Greek - suffix with secondary lengthening - subwths, protostaths.
It was only later that the first of those suffixes succeeded, and gave nouns like praedicator or benefactor, which are newer.
This is what we have now to translate those common words into Indo-European and to correct our works. If you have some comment related to these questions, please share!