There is no lack of theories on morphological conversion (Bauer, Lieber and Plag 2013:563 ff). Some of the oldest theories used the concept of zero affixation, i.e. the postulation of an affix without phonological content that is responsible for the change in category and meaning. Marchand (1969:359 ff) gives an historical foundation for zero derivation: Following Jespersen, he suggests that the loss of endings gave rise to derivation by a zero morpheme. Synchronically, however, there are various problems with the approach: It is impossible to decide whether it is a zero prefix or a zero suffix, zeroes tend to proliferate (one for deverbal nouns of common gender, another one for neuter nouns, one for denominal verbs, one for de-adjectival verbs, etc.), and it is hard to imagine how language learners could ever learn to master the mechanisms.

• The direction of conversion is not always easy to decide. This is especially true for noun-verb pairs, adjectives-adverb pairs, and preposition-adverb pairs in Afrikaans.
• One way to determine the direction of conversion, is to consider it diachronically: Which function did the word have first?
  • The proper noun Google existed before the converted verb google. This is therefore a clear example of verbalising (or N-to-V) conversion.
  • With regard to an N-V pair like wals waltz and (om te) wals (to) waltz, one can consider the etymology of the word. Wijk (1936 [1912]) shows that the High German noun walzer waltz was derived from the verb walzen to turn, roll, and we can therefore conclude that – diachronically speaking – the noun was converted from the verb.
• Another way to determine the direction of conversion, is to consider it synchronically: Which function does the word have primarily?
  • Primacy can be determined in various ways, including distribution, (perceived) frequency, prominence, definitional properties, etc.
  • From a corpus investigation it is clear that the verb teks to text occurs much less frequent than its noun counterpart. Since this verbal function is a new development, one can safely assume that the noun was converted into a verb.
  • In the case of a very old word like vinnig quick(ly); fast(ly) one should tread carefully, since vinnig is used more frequently as an adverb than as an adjective. Despite the fact that vinnig originated as an adjective, it is now used more frequently as an adverb. In such a case the adverbial use is primary, but the conversion was still from ADJ to ADV.
  • One could ask users to define a word pair, and the function that is used in the definition of one of the words, could be considered the primary function. For example, one could ask users in a questionnaire to define the verb and noun dokter doctor. If most users define the verb as to act like a doctor, then the nominal function is primary.
  • Similarly, one could consider definitions in dictionaries. For example, WAT defines the verb ploeg plough as to cultivate land with a plough, with a cross-reference to the noun ploeg plough. We can then deduct that the directionality of the conversion is from N to V.
  • Oftentimes, dictionaries use primacy as an ordering principle for the sequence in which functions and polysemous meanings are presented. For example, WAT presents the nominal function under the lemma haak (to) hook first, followed by the verbal function, but vice versa under the lemma kap (to) cut. One could then deduce N-to-V conversion in then case of haak, and V-to-N conversion in the case of kap.
  • Kempen (1969:37-38, 40) also illustrates that the number of polysemous meanings for the different functions of a word could serve as an indication of the primary function. For example, in WAT the noun baan lane; circuit; etc. has 19 polysemous meanings, while the verb has only two. From this we can deduce that the noun is primary, and hence that the verb was formed through N-to-V conversion.

The notion conversion might be extended to cases where there is no change in category, but in subcategory. For instance, it is often possible to use a causative verb formed by means of the causative affixes ver- or -iseer as an intransitive verb as well:

Another type of category-internal valency change is the formation of middle verbs on the basis of verbs that occur with an object or a prepositional adjunct that denotes an instrument, a location, or an external circumstance:

As argued in Booij (1992) and Ackema and Schoorlemmer (1994;1995), this intransitive use of verbs cannot be a matter of syntax, but requires a lexical rule that applies to the lexical conceptual structure of verbs that express an action: the resulting middle verbs do not express an action, but a property. These middle verbs usually require some evaluative expression to be present in the clause in which they occur that specifies how well the subject can be involved in the event.

The three tables below have been reproduced from Theron's (1974) thesis.

Words functioning primarily as verbs, and secondarily as nouns: