NB: This sub-section discusses Velar Harmony which is a particular type of Consonant Harmony.

Velar harmony process

Alveolars are particularly susceptible to harmony. Here is another example.






Again, we see that a target alveolar has assimilated to the place of articulation of a trigger velar, i.e. the voiceless alveolar plosive /t/ has assimilated to the voiceless velar plosive /k/. Now, because the affected segment (target) precedes the segment causing the assimilation (trigger) this is an example of regressive assimilation (see Table 13 in Assimilation). In other words, the harmony spreads from right to left (Figure 9).

Figure 9. Regressive velar harmony affecting an alveolar

Figure 9. Regressive velar harmony affecting an alveolar.

Here are some more examples of regressive velar harmony affecting alveolars:















This type of velar harmony is evident in typically developing speech up to about 2;04 years. In child language, there are also examples of progressive velar harmony affecting alveolars, i.e. with the harmony spreading from left to right:















This progressive process is less persistent than its regressive counterpart. Whereas regressive velar harmony affecting alveolars is exhibited until about 2;04 years, progressive velar harmony affecting alveolars has usually stopped by around 1;11 years of age.

Now, as well as velar harmony targeting alveolars, it can also target bilabials, e.g.






We see here that the target bilabial /b/ has assimilated to the place of articulation of a trigger velar /g/, i.e. the voiced bilabial plosive /b/ has assimilated to the voiced velar plosive /g/. Once more, because the target precedes the trigger this is another example of regressive assimilation, with the harmony spreading from right to left (Figure 10).

Figure 10. Regressive velar harmony affecting a bilabial

Figure 10. Regressive velar harmony affecting a bilabial.

The following are further examples of regressive velar harmony affecting bilabials:















This type of velar harmony can be seen in typically developing speech up to about 2;00 years of age. Whilst there are examples of progressive velar harmony affecting bilabials, this process is usually eradicated around 1;06 years of age. Here are some examples:















Confusion with backing

Whilst getting to grips with the concept of velar harmony, you may have noticed that all of the examples provided in this subsection are also possible examples of the phonological process of backing discussed elsewhere. Let us reconsider our first example:






Here, the target consonant /t/ is realized as /k/. Now, we know that backing is defined as replacing a non-velar or non-glottal consonant with a velar or glottal consonant. Therefore, is it not arguable that – because the initial non-velar and non-glottal /t/ has been replaced with a velar – this is, in fact, an example of backing? How, then, do we determine whether or not a child realizing the word tick as /kɪk/ is, indeed, an example of velar harmony or backing?

We must bear in mind that if we were assessing a person’s phonological system then we would not be making decisions about which underlying processes are operating based solely on the analysis of just one word. In order to identify velar harmony, we need to investigate whether the affected consonant changes to a velar in contexts lacking a neighboring velar.

Remember that velar harmony is an assimilatory process that requires BOTH a target and a trigger consonant, i.e. the context must always minimally be CVC. This is the situation with the word tick, which is a good example of a CVC word. Now, recall that backing is a systemic substitution which can operate in different contexts, e.g. in CV words and VC word. So, in order to identify
whether or not a consonant is subject to consonant harmony or backing we need to examine how it is affected in contexts such as these, i.e. lacking a neighboring velar consonant. Figure 11 sets
out some possibilities.

Figure 11. Velar harmony or backing

Figure 11. Velar harmony or backing?

None of these test words contains a velar. So, if each test word is realized as the appropriate adult form then there appears to be no process affecting the consonant /t/. It is not being backed and it cannot undergo velar harmony because there is no context in which the other consonant in a CVC structure is a velar. However, if /t/ were realized as /k/ in any of the test words then this would be backing. So, considering the test items above, if the word tick is realized as /kɪk/ but the words tea, eat and tape are realized as the appropriate adult forms, then we can (tentatively) state that the underlying process operating here is consonant harmony and not backing.

In summary, velar harmony may be confused with backing and investigators need to determine whether or not the target consonant changes to a velar in contexts lacking a velar.


All normative data for consonant harmony in this section is taken from Pater & Werle (2003). 

Pater, J., & Werle, A. (2003). Direction of assimilation in child consonant harmony. The Canadian Journal of Linguistics , 48 (3/4), 225-238.

NEXT>> Labial Harmony