The typology of case relators refers primarily to two of their grammatical properties:

1. their degree of grammaticalization, on a scale from complex adposition to simple adposition to agglutinated case affix to fusional case inflection
2. their sequential order, more precisely their position relative to the dependent nominal that they govern.

The second criterion may be visualized more completely by the following diagram:

While the traditional terms of 'regens ante rectum' and 'rectum ante regens' are entirely appropriate in this domain, they do not generalize to attribution and adjunction as these are relations of modification, not of government. If one wants a word order typology that comprises all of the syntactic relations, the terms are left-branching (including properly rectum ante regens) and right-branching (including properly regens ante rectum).

Turning now to criterion #1, the stages of this grammaticalization path differ according to the relevant parameters of grammaticalization. They are illustrated by the following example series of local case relators:

(repeated from another section) shows a complex postposition composed of a region noun and a simple postposition. shows a simple postposition, this being enclitic in the language. shows an agglutinative case suffix and finally shows case inflection.

An adposition differs from a case affix by its structural scope: while a case affix typically attaches to a noun, an adposition may govern an entire NP. This may be seen in : in the syntactic construction underlying the univerbation to a complex postposition, the postposition na̱ (inessive) governs the preceding noun phrase. Criteria enabling us to apply the distinction in particular cases depend on the language. In Turkish, case suffixes as in are subject to vowel harmony, postpositions are not. The fusional case suffix in differs from an agglutinative one by being inseparable. On the one hand, neither the root dom- nor the stem domu- are possible word forms. On the other, the suffix -i is an exponent of more than one morphological category, viz. number and case, and its allomorphy depends on the thematic vowel of the stem and its gender.

It has generally been observed that postpositions grammaticalize easily to case suffixes, while prepositions grammaticalize seldom to case prefixes. The explanation seems to lie with the word order inside the noun phrase. Subconstituents of a noun phrase are mainly prenominal, even in languages which otherwise have right-branching word order. Languages with postpositions are generally predominantly left-branching, inside the NP just as above its level. As a consequence, the postposition always follows directly the governed noun, as it does in f. Given this fixed combination, coalescence is favored. By contrast, a preposition does not generally directly precede the governed noun because of the widespread tendency of left-branching constituency inside the NP. The French NP is a case in point. The preposition de (gen) is maximally grammaticalized. It directly precedes the governed noun in a, but the adjective attribute in #b and the definite article in #c.

The fact that the preposition does merge with the definite article shows that it would develop into a case prefix were it not for the varying coconstituency in the prepositional phrase.

Not all languages have adpositions. Yidiny (Pama-Nyungan, Australia) has none. It is also possible for a language, e.g. Ancient Chinese, to convey case relations only by nouns or verbs which are not grammaticalized to adpositions. The majority of the languages of the world does have adpositions. These differ by the degree of grammaticalization exhibited by their paradigms of case relators. The type illustrated by f, viz. a language with complex and simple adpositions, is frequent, both with prepositions and with postpositions. Many languages of this latter type agglutinate simple postpositions to the noun and then have a paradigm of case suffixes, too. This is the situation in Turkish.

The situation represented by Latin is noteworthy. The language has case inflection in a closed and small paradigm of at most seven cases as illustrated by . With the exception of two isolated denominal postpositions, all more specific case relations are coded by simple prepositions. The suffixal case inflection is inherited from Proto-Indo-European. A paradigm of case suffixes can be enriched and renewed by postpositions, as is observable in Turkish. However, at the prehistorical stage where Indo-European languages were drawing their adpositions from Proto-Indo-European adverbs, Latin was already embarking on the way to right-branching syntax (only completed in the Romance period) and opted for prepositions. This was a fatal decision, since it dried out the source for new case suffixes. As a result in Proto-Romance, the grammaticalization of the case-suffix system ended up in loss, while at the same time the prepositions were grammaticalized to case markers.