Introduction. While each language has a certain inventory of basic sounds, there also seems to be a set of rules of how those sounds can be combined within the language. Some words just sound right to native speakers, while others sound strange or foreign. Phonology studies these patterns in language at several levels. The word is the basic, isolatable unit of language. Words can be divided into syllables, and syllables into segments. Segments can be further analyzed based on their features. In Biblical Hebrew the rules of shewa are an example of phonological rules.
1. Segments. Segments are linear divisions of the speech stream at the level below syllables. We are used to dividing speech into segments when we spell words phonetically, for the most part a letter represents a segment. Segments are sometimes described as the smallest sound unit that can be used to distinguish meaning. A segment is said to contrast (or be in distinction or opposition) when its presence alone can distinguish meaning between two different words.
Demonstrating which segments contrast for a given language uses the method of minimal pairs. A minimal pair is a pair of words which have different meanings but only differ by one segment. For instance, in English the words [sip] and [zip] show that the segments [s] and [z] contrast.
1.1. Phonetic variation If you listen closely, there are many more phonetic sounds used in everyday speech then the set of contrasting segments established for the language. Much of this phonetic variation is systematic and conditioned by the surrounding phonetic environment. As mentioned before, speech is a complex action involving multiple organs which must all be in sync. In order to articulate more efficiently, we often anticipate the next point of articulation which can cause slight variations in how we pronounce a given segment. The larger phonological context such as the prosody can also cause phonetic variation. Because the variations are largely regular, native speakers are able to filter them out and focus only on the relevant contrasts. Cases where phonetic variation occurs within the same sound environment are called free variation.
1.2. Phonemes and Allophones. Native speakers seem to have the ability to group the predictable phonetic variations of a contrasting segment into a class of sounds labeled on the phonological level as a phoneme. Each of the phonetic variants are called allophones. A large part of phonology is to analyze the phonemes of a language and their allophonic variation. For instance, few English speakers notice that the /s/ is pronounced differently in the words “cats” and “dogs”. In the second case it sounds more like [z] because of the influence of the [g]. Note that in Linguistics works phonemes and phonemic representations are often placed between backslashes (/s/) while phonetic transcriptions are placed between brackets ([s] or [z]).
An example of allophonic variation in Hebrew are the so-called בגדכפת letters. Depending upon the where they occur in a syllable, these letters may be pronounced either “hard” or “soft”. For instance, [b] can be pronounced either hard /b/ or soft /v/. This change in pronunciation does not change the meaning of the word (though notice how the soft pronunciation may overlap with the contrasting phoneme [v]).
2. Features. Features are the phonological unit below the segment. To describe a segment as a combination of features allows the linguist to analyze the various independent activities that are combined in the articulation of a given segment and to classify sounds based on these separate phonological characteristics. Phonological rules governing how sounds can be combined and when allophonic variation occurs can then be created at the feature level instead of creating a rule for every individual segment.
Features are written in a square bracket such as [syllabic] or [consonantal]. A given segment is then described as a binary array of features where a feature is marked with a plus [+] if it is present and a minus [-] if it is absent. For example, the vowel [a] can be described as [+syllabic, -consonantal, +sonorant, -high, + low, +back] (Note that different linguists may use different sets of features).
For example, in English many words start with /sp/, /sk/, or /st/, but never /sb/, /sg/, or /sd/. What is the difference? In the first set, the second consonants are all voiceless [-voice] while in the second they are voiced [+voice].
3. Syllables. The syllable is the phonological unit between the segment and the word. It is usually described as a nucleus (or peak) surrounded by associated non-syllabic segments. The nucleus is a highly sonorant element, either a vowel or a sonorant consonant such as a liquid. Usually the syllable is broken into two parts: the onset and the rhyme. The onset is comprised of the non-syllabic segments preceding the nucleus. The rhyme is comprised of the nucleus and the coda – the non-syllabic segments following the nucleus. For instance, a syllabic analysis of the word “spin” gives:
|Onset (O)||Rhyme (R)|
|Nucleus (N)||Coda (C)|
Technically, the nucleus is the only required element of a syllable. While there do seem to be some universal principles for forming syllables, each language also has specific constraints on what constitutes a valid onset or coda.
Most syllables follow the sonority heirarchy moving from less sonorant to more sonorant segments towards the nucleus and the reverse away from it. Vowels are the most sonorous followed by the glides, liquids, nasals, fricatives and stops. However, individual languages allow exceptions. For instance, many languages allow the fricative /s/ to begin (or end) a consonant cluster even if it is followed by a stop such as the sequence /s/+/p/ in the onset of /spin/.