File 4.4, 2.2-2.5
Read this => Because some of the phonetic symbols used in these problems are unavailable in this format, I have used substitutes. These are explained below, where relevant.
Problem 2.2 Spanish
Here, [G] represents the voiced velar fricative and [B] represents the voiced bilabial fricative.
Question [a]: The sounds [G] and [g] are allophones of a single phoneme. They are in complementary distribution, as follows:
The sounds [b] and [B] are also allophones of a single phoneme. They are in complementary distribution, as follows.
What is striking here is that in both cases, the fricative allophone occurs after vowels, while the stop allophone occurs everywhere else. This suggests strongly that a more general process is involved which applies not just to [b]/[B] and [g]/[G], but to all voiced stop/fricative pairs.
This is confirmed when we look at the data from the earlier problem on [d]/D]. Question [b]:
This is confirmed when we look at the data from the earlier problem on [d]/D]. Thus we can say more generally that all voiced obstruents have two allophones: a voiced fricative which occurs after vowels and a voiced stop which occurs elsewhere.
Note: Since the voiced stop is the "elsewhere" case (the one that occurs in more different environments), we regard that allophone as the basic one, with the fricative derived from it. That means that in the phonemic (underlying, mental) representation, only the voiced stops occur. The fricatives occur only lin the phonetic representation and are derived by the following rule in Spanish:
If the rule does not apply, then the sound is pronounced as a voiced stop.
Note: There are two sounds here, the voiceless alveolar stop [t] and the voiceless alveolar affricate, which I represent as [ts].
The two sounds are distributed as follows:
The two sounds never occur before the same vowels. Thus, they are in complementary distribution and are allophones of the same phoneme.
It looks like [t] is the basic allophone because it occurs before a wider range of vowels. The class of vowels that it precedes does not form a natural class, while the class of sounds that [ts] precedes does form a natural class. Thus it is easier to take [t] as basic and write a simple rule to derive [ts] from it than the other way around. Therefore, only /t/ occurs in the phonemic representation. [ts] is derived from it by the following rule.
If the rule does not apply, then the phoneme is pronounced [t].
Note: I am using [c] here to represent the voiceless palatal fricative.
[x] and [c] are in complementary distribution in German. [x] occurs only after back vowels and [c] occurs only after front vowels. Thus, they are allophones of the same phoneme.
It is not obvious which is the basic allophone, because each sound occurs in a simple phonetic context, one which can be described in terms of a natural class of sounds. However, I will assume the basic allophone is [x]. This means that in phonemic representations, only /x/ occurs, and [c] is derived by the phonological rule below. If the rule does not apply, then /x/ is pronounced [x]:
Note: This is an assimilation rule, since the consonant is articulated further forward in the mouth after a front vowel.
There are three sounds here, which I represent as follows:
The basic allophone is the voiced trill, [R]. This is because the phonetic environments in which it occurs cannot be described as a natural class. It would therefore be very awkward to pick one of the other sounds as the basic allophone and write a rule to turn it into [R]. It is easier to take [R] as the basic allophone and write rules to turn it into each of the other sounds.
Hence, in the phonemic representation, only /R/ occurs. Farsi has two phonological rules which convert the phonemic representation into a phonetic one:
If neither rule applies, then /R/ is pronounced [R], the voiced trill.