The allophonic rules of Spanish

1. The basic concept
According to the tables on the phonemes page, the standard European variety of Spanish has a total of 24 phonemes (5 vowels and 19 consonants). On the other hand, the relevant tables on the vowels and consonants pages attribute a combined total of 35 distinct speech sounds (5 vowels, 28 consonants and 2 semivowels) to the same variety. The reason for this mismatch is that phonemes are essentially classes of sounds and several phonemes in Spanish have more than one member.

2. The phonemes /b, d, g/
This point can be illustrated first of all with the voiced stops [b, d, g] and the voiced approximants [β, ð̞, ɣ]. A parallel exists between these two groups of sounds, in that the places of articulation of the stops exactly match those of the approximants, as shown in Table 1 below.

Table 1 Voiced stops and approximants in Spanish
	Stop	Approximant
Bilabial	[b]	[β]
Dental	[d]	[ð̞]
Velar	[g]	[ɣ]

It turns out that there are no cases of minimal contrast between a voiced stop and the corresponding voiced approximant. This is because in each case the stop and the corresponding approximant are in complementary distribution, i.e. they occur in mutually exclusive contexts. Thus the stops are restricted to the following contexts:

initial position: [ˈbino] vino ‘wine’, [ˈdoj] doy ‘I give’, [ˈgoma] goma ‘rubber’
after a nasal consonant: [ˈambos] ambos ‘both’, [ˈan̪do] ando ‘I walk’, [ˈteŋgo] tengo ‘I have’
in the case of [d], after [l]: [ˈtoldo] toldo ‘awning’

The approximants, in contrast, have the following distribution:

after a vowel: [aˈβeɾ] haber ‘to have’, [ˈloð̞o] lodo ‘mud’, [ˈlaɣo] lago ‘lake’
after consonants other than nasals or, in the case of [ð̞], laterals: [ˈalβa] alba ‘dawn’, [ˈaɾð̞e] arde ‘burns’, [ˈalɣo] algo ‘something’

Because they are in complementary distribution, a voiced stop and its corresponding approximant can be regarded as contextual variants or allophones of the same phoneme. In other words [b] and [β] are allophones of a single phoneme (conventionally represented as /b/), [d] and [ð̞] are allophones of a single phoneme (conventionally represented as /d/) and [g] and [ɣ] are allophones of a single phoneme (conventionally represented as /g/).

3. The coronal fricatives /θ/ and /s/
The dental fricative phoneme /θ/ and the alveolar one /s/ are also subject to an allophonic rule, both being voiced before an immediately following voiced consonant. In the case of /s/, the resultant allophone is [z], the voiced alveolar fricative, and in the case of /θ/ the resultant allophone is [ð], the fricative counterpart to the voiced dental approximant [ð̞] discussed above:

/θ, s/ = [ð, z] (before voiced consonant):
[xuðˈɣaɾ] juzgar ‘to judge’
[ˈmizmo] mismo ‘same’

/θ, s/ = [θ, s] (all other contexts):
[koˈθina] cocina ‘kitchen’
[ˈkasa] casa ‘house’

The voiced dental fricative [ð], an allophone of /θ/, is similar in its articulation to the dental approximant [ð̞], which is an allophone of /d/. The main difference is that [ð̞] is pronounced with little or no audible friction while [ð] does exhibit audible friction. Interestingly, this latter sound was originally an allophone of /d/ rather than /θ/, as is indicated by the fact that words which now have the letter z before a voiced consonant such as /g/ or /m/ were previously spelled with d. For example, juzgar, used above to illustrate the voicing of preconsonantal /θ/, was originally spelled iudgar, implying that the preconsonantal [ð] in this word was analysed by medieval speakers as being an allophone of /d/ (at the time /θ/ had not yet come into existence). Analogously, the words now spelled hallazgo ‘finding’ and bizma ‘poultice’ were written in Old Spanish as falladgo and bidma respectively. This topic is discussed in depth in Mackenzie 2022 (see especially Section 3).

4. The voiced palatal phoneme /ʝ/
The palatal phoneme /ʝ/ is ‘hardened’ into an affricate in certain positions. More specifically, it is realized as the voiced palatal affricate [ɟʝ] in initial position and after a nasal. In all other contexts it is articulated as a voiced fricative, notated by the symbol [ʝ]:

/ʝ/ = [ɟʝ] (initial position and after nasal):
[ɟʝo] yo ‘I/me’
[ˈkoɲɟʝuxe] cónyuge ‘spouse’

/ʝ/ = [ʝ] (all other contexts):
[aˈʝeɾ] ayer ‘yesterday’

For n = [ɲ] in cónyuge, see Section 5 immediately below.

5. The nasal consonant /n/
When immediately followed by another consonant, /n/ assimilates – i.e. becomes more similar – to that consonant in terms of its place of articulation. Thus /n/ is realized as as labiodental [ɱ] before /f/, as dental [n̪] before /θ, t, d/ and as velar [ŋ] before /k, g, x/:

[eɱˈfeɾmo] enfermo ‘ill’
[ˈben̪da] venda ‘bandage’
[ˈbeŋga] venga ‘come on’

Notice also that in the orthographic sequence ny, the letter n stands for the palatal nasal phoneme /ɲ/ – the one normally represented by the letter ñ – rather than /n/. This occurs, for example, in cónyuge, which is transcribed in Section 4 above.

6. The semivowels [j] and [w]
Finally, the semivowels [j] and [w] can be assigned to the phonemes /i/ and /u/, as there are no minimal contrasts in Spanish between the sounds [i] and [j] or between [u] and [w]. Thus the semivowels can be seen as the forms taken by /i/ or /u/ when they occur in a diphthong with another vowel. Each semivowel can occur immediately before or immediately after the vowel with which it forms a diphthong, as is shown in Table 2 below.

Table 2 Distribution of the semivowels [j] and [w]
	Before vowel	After vowel
[j]	[ˈtjera] tierra ‘land’	[ˈbojna] boina ‘beret’
[w]	[ˈfweɣo] fuego ‘fire’	[ewˈɾopa] Europa ‘Europe’