With geometrical, dimensional views on the phonemes, it's reasonable to aspect that the choice of phonemes for different semantic and syntactic use isn't totally random.
   Naturally, there are a lot of other factors that differentiate languages as the fact that language is inherited, further sound shift "laws", different environments and cultures etc.
   The geometrical aspects alone, with "dimension chains" of both levels of speech, of phoneme types and of articulation positions of phonemes, give also a multitude of possibilities. There are, however, some interesting examples from the literature that may reveal the impact of these geometries on choice of phonemes.

- m, n, l, r represents ordinary noun suffixes in both Indo-European and Semitic languages (LB).
- m - n - l - t seems common in Swahili as phonemes in affixes according to a phrase book.

Types - Positions, connections in the loop model:

- n and l are common in case endings in Hungarian, for example:
n a case ending with the sense "in, inside", l a suffix for "out of or of" (BC). (It seems to mirror positions!)

Grammar (syntax) as d-degree 4 expressed in case endings and affixes: Compare step 4-3 in the big underlying chain for levels in speech and nasals in step 4-3:

- As for phoneme types it looks here as plosives and their fricative versions should constitute the body in morphemes or word stems ?
   If phoneme types P and F through some deep disposition should be chosen for the central body of consonants in word stems, it could perhaps have connection with the same steps in the big level chain, where we suggested that word classes (here = "parts of sentences") developed in step 3-2, morphemes in step 2-1.

Both nasals and liquids as affixes are characterized by a branched air current, but in complementary ways: for nasals depending on soft palate with uvula from above (dorsal side) and outer stop (as inwards defined), for liquids on the tongue raised from below (ventral side); a correspondence uvula - tongue could be noted. The air current is branched vertically in nasals, horizontally in liquids, also a complementarity.
   Compare perhaps syntactic case endings as debranched to the level of word classes - and phonetically liquids in step 1→0/00 as debranched from higher d-degree steps.
5 → 4 → 3 →...

In many of the examples that Linus Brunner gives (LB p. 44) on the reversal of the 3rd nuancing consonants in Semitic word stems to 2nd position in Indo-European languages, this nuancing, differentiating consonant is n, (m), r and l. If appearing as 2nd consonant in the IE-languages, they could therefore be original endings.

The sequence of phonemes in words becomes a change in direction, analogous to the development of step 1-0/00 inwards step 3 - 2 in our dimension chain: l and n in the examples as a kind of infix:

The reversal of such phonemes were fully implemented only in the verbs (LB). Why? Perhaps because reversals imply switching direction, and verbs primarily may be connected with directions and vectors - as we have done in these interpretations?

Regarding the dimension chain for positions, the outer nasal n are formed approximately at the same position as the front plosives d and t and front l-sound and represents a lower d-degree step, belongs in positions to "2 - 1"-phonemes.
   Among Uralic languages, it's said (BC) that a native Finnish word stem or smallest, independent, sense-carrying morphemes (what BC calls a link) can only end in consonants n, t, s, r,. l, that's positions "2 - 1", AP-A-DA.
   (Cf. our interpretation of the agglutinating language type as corresponding to a loop: steps 4 —> 3 ...... 2 <—1.)

The l-sound represents unfinished action in Finnish. It's also used in reflexive meaning (BC): that's direction d-degree 1 as inwards.
   Again, d-degree 1 - eventually polarized to 0/00 of motions, is represented in each step in the dimension chain, outwards or inwards This could explain why the liquids l and r are freely exchangeable in many languages.
.
In such things as reflexive pronouns, (as in words as round and rotation for instance), the r-phoneme could illustrate the turn of direction in last d-degree 0/00 of motions, the referring back, ~ inwards in the dimension chain.

Outwards, the r-phoneme, the vibrant, occurs often in "frequentive" verbs in Samoyedic, that is verbs with sense of repetition. Compare in Swedish r-suffix for indefinite plural, ~ 00-pole. There is also the r-sound (in both Swedish and English in comparatives, a kind of plural or repetition as enforcement of adjectives.
   The vibrant, the r-sound in d-degree 0/00 of motions in our dimension model, is often used as such, in Swedish as verb ending (singular, in modern Swedish also plural) for present tense (-er). The r-sound is specially frequent also as suffixes of verb stems in Semitic languages. Hence, verbs here regarded from the aspect of activity, of motions.
   Already Plato wrote about the r-sound in words for motions, according to a quotation (BC).

In connection with word categories as verbs, nouns, adjectives, there are in Indo-European languages certain differences between nasals and liquids as phonemes in suffixes (LB):

(Aspects in file Word classes.)

m, n are foremost suffixes for nouns and adjectives. The lateral l represent often a diminutive form and tools ("endings" or "suffixes" of the human hand !), as in some Uralic languages (BC). The vibrant r occur most frequently as suffix in adjectives.
   Thus, we can find a certain parallel between phonemes and word categories in the displacement outwards in a dimension chain: nasals —> liquids, a shift towards outer phoneme types, and from nouns to nouns in secondary roles to adjectives, according to the interpretation of the dimension chain for word categories..

In Uralic languages, p and k become endings for nouns derived from verbs (BC, u), which we could interpret as result of a feeling that plosives as total barriers are connected with nouns as closed units?
   Other derivative endings according to the same reference, as pt or kt for verbs derived from verbs, contradict obviously any unambiguous interpretation. Could they perhaps hide an intermediate substantivized phase of the verb? (As in form → formula → formalize.)
   Alternatively, one could eventually see traces of the verb as direction in the movement of the tongue from lower to upper level of plosives in p → t, which also is a movement from inner to outer plosive in k→t ?

The r-sound became a case suffix in nominative - accusative in (many?) Indo-European languages, according to one statement. This could perhaps be derived from a feeling that subject - object in a sentence represented a centre - anticentre (0 — 00) relation, the two phases or parts 0/00 in the vibrant?

The phonemes n and t are endings for different classes of nouns in definite form in Swedish, eventually forms that first reflected personal versus neuter nouns (?), now randomly distributed among all kind of things. Cf. phonemes n and t with internal/external, with living/lifeless, voiced versus voiceless sounds: ( "en - ett" = one).

In vowels we have the strong verbs were past tense often is marked by a deeper or back vowel, an inward direction.

These few examples and speculations about interpretations demonstrate mostly that phonemes don't need to be arbitrarily selected ("initially"), that we can assume a link between phonemes and meaning, the start and end in the level chain of language.
   It implies that development of languages may have been "onomatopoeic" in a much wider sense than the ordinary imitating one: speech organs used to "imitate" or illustrate the deeper, complex geometries of own and external Nature.

*

To Sound Shifts