You are viewing a free preview of this lesson.
Subscribe to unlock all 10 lessons in this course and every other course on LearningBro.
Phonology is not simply a catalogue of the sounds in a language — it is the study of the rules and patterns that govern how those sounds behave. Every language has a system that determines which sounds can occur in which positions, how sounds interact with their neighbours, and how the abstract phonemic representation stored in a speaker's mind is converted into the actual phonetic sounds that leave the mouth. For AQA A-Level English Language (7702), this is not an isolated topic but a method integrated across every component — Paper 1, Paper 2 and the NEA. Understanding phonological processes lets you move beyond merely labelling an accent feature to explaining the systematic rule that produces it, which is exactly the depth that lifts analysis from description into evidenced argument. This precision serves AO1 (systematic application of method and terminology) and, when you connect a process to its social meaning, AO3 (contextual factors). The full assessment-objective profile for the A-Level is AO1 26 · AO2 26 · AO3 23 · AO4 15 · AO5 10.
The organising idea behind the whole lesson is the distinction between phoneme and allophone, and between the underlying form a word "has" and the surface form it is realised as. A phonological process is, in essence, a regular relationship between those two levels: given the underlying /t/ of "water", the process of t-glottalling derives the surface [ʔ] in the right environment. Keeping the two levels straight — and signalling which you mean with slashes / / versus square brackets [ ] — is the single most reliable marker of phonological competence on this course.
The concept of the minimal pair is the foundational tool of phonemic analysis. A minimal pair is two words that differ in meaning but are identical except for a single sound in the same position. If swapping one sound for another yields a different word, the two sounds must belong to different phonemes, because the difference between them is doing meaningful work.
| Word 1 | Word 2 | Contrast | Conclusion |
|---|---|---|---|
pat /pæt/ | bat /bæt/ | /p/ vs /b/ | /p/ and /b/ are separate phonemes |
sit /sɪt/ | set /set/ | /ɪ/ vs /e/ | /ɪ/ and /e/ are separate phonemes |
thin /θɪn/ | fin /fɪn/ | /θ/ vs /f/ | /θ/ and /f/ are separate phonemes (in RP) |
sing /sɪŋ/ | sin /sɪn/ | /ŋ/ vs /n/ | /ŋ/ and /n/ are separate phonemes |
bat /bæt/ | bad /bæd/ | /t/ vs /d/ | /t/ and /d/ are separate phonemes |
The logic is decisive: a single minimal pair is sufficient to prove phonemic status, because it shows the contrast is exploited to distinguish meaning. This is why the minimal-pair test is the gold standard for establishing the phoneme inventory of an accent.
Key Definition: Minimal pair — two words identical in every respect except for one sound in the same position, demonstrating that the two sounds are separate phonemes. For example, "pin"
/pɪn/and "bin"/bɪn/prove/p/and/b/are distinct phonemes of English.
Sometimes a perfect minimal pair is hard to find — /ʒ/ is rare and almost never word-initial — so a near-minimal pair provides supporting evidence: two words differing in the target sounds and very similar (though not identical) elsewhere. "Measure" /ˈmeʒə/ beside "mesher" /ˈmeʃə/ is a near-minimal pair supporting the /ʒ/–/ʃ/ contrast; "azure" beside "assure" works similarly. Near-minimal pairs are a useful fallback when the phoneme you are investigating is distributionally restricted.
A crucial subtlety for A-Level: minimal pairs are defined relative to an accent. "Thin"/"fin" is a minimal pair for a speaker who keeps /θ/ and /f/ apart — but for a TH-fronting speaker who realises both as /f/, the two words are homophones /fɪn/ and the contrast has collapsed. Likewise "put"/"putt" is a minimal pair in RP but not in an unsplit northern accent where both are /pʊt/. So the same word-pair can be a minimal pair in one accent and not in another, which is precisely how mergers and splits reshape a phoneme inventory over time.
Two sounds are in complementary distribution when they never occur in the same phonetic environment — each appears exactly where the other cannot. This distributional pattern is the signature of allophones of a single phoneme: because the two never compete in the same slot, a speaker could never use the difference between them to signal a change of meaning. The choice between them is predictable from context.
The classic English example is clear versus dark /l/:
| Allophone | Environment | Examples |
|---|---|---|
Clear [l] | Before a vowel (syllable onset) | light [laɪt], play, alive |
Dark [ɫ] | After a vowel / before a consonant (syllable coda) | fill [fɪɫ], milk, people |
Clear [l] and dark [ɫ] never appear in the same position, so they are in complementary distribution and are therefore allophones of the single phoneme /l/. A speaker could use either anywhere without changing meaning — it would merely sound odd. The same pattern governs aspiration:
| Allophone | Environment | Examples |
|---|---|---|
Aspirated [pʰ tʰ kʰ] | Start of a stressed syllable | [pʰɪn], [tʰɒp], [kʰæt] |
Unaspirated [p t k] | After /s/ | [spɪn], [stɒp], [skæt] |
You can verify aspiration physically: a hand held before the mouth feels the puff of air on "pin" but not on "spin". Because the aspirated and unaspirated variants are in complementary distribution, they are allophones of /p t k/, not separate phonemes.
Key Definition: Complementary distribution — a pattern in which two sounds never occur in the same phonetic environment, each appearing where the other cannot. Sounds in complementary distribution are allophones of one phoneme, not separate phonemes.
The two diagnostics work as a neat decision procedure. Can the two sounds form a minimal pair? If yes, they are separate phonemes. Are they in complementary distribution (predictable from environment)? If yes, they are allophones of one phoneme. This is the central analytical routine of phonemic analysis, and being able to apply it explicitly to data is a strong AO1 discriminator.
Two sounds are in free variation when they can occur in the same environment without changing meaning, and the choice is not fixed by phonetic context. Unlike complementary distribution (where the environment dictates which allophone appears), free variants are interchangeable in the same slot.
| Variant 1 | Variant 2 | Example | Environment |
|---|---|---|---|
[t] (alveolar plosive) | [ʔ] (glottal stop) | "butter" — [ˈbʌtə] or [ˈbʌʔə] | Intervocalic /t/ in many British accents |
[ɹ] (post-alveolar approximant) | [ɾ] (alveolar tap) | "very" — [ˈveɹi] or [ˈveɾi] | /r/ in some accents |
[ɪŋ] | [ɪn] | "running" — [ˈrʌnɪŋ] or [ˈrʌnɪn] | Word-final <ing> |
Strictly, much apparent "free" variation is not random at all but socially conditioned: the choice between [ɪŋ] and [ɪn], or between [t] and [ʔ], correlates with social class, age, sex and formality. When variation patterns systematically with social factors in this way, sociolinguists call the alternating forms a sociolinguistic variable (notated with parentheses, e.g. (ng), (t)) rather than treating the variants as truly "free". This is the bridge between the phonological notion of free variation and the sociolinguistic studies of Lesson 10: the same alternation that looks "free" from a purely structural viewpoint turns out, when quantified across speakers, to be a finely patterned index of social identity.
Key Definition: Free variation — a pattern in which two or more sounds can occur in the same phonetic environment without changing meaning, with the choice not determined by phonetic context. Much free variation is in fact socially conditioned, making the variants sociolinguistic variables.
Phonotactics is the set of rules governing which sequences of sounds a language permits. Every language constrains how phonemes may combine, and English has particularly rich constraints on syllable onsets and codas.
| Position | Rule | Permitted | Disallowed |
|---|---|---|---|
| Onset | Up to 3 consonants; any 3-consonant onset must begin with /s/ | /str/ (string), /spl/ (splash), /skw/ (square) | */tl/, */dl/ as onsets |
| Onset | After /s/, only voiceless plosives /p t k/ may follow | /sp/ (spin), /st/ (stop), /sk/ (skip) | */sb/, */sd/, */sɡ/ |
| Onset | Two-consonant onsets follow fixed patterns | /pl/ (play), /pr/ (pray), /tr/ (try) | */pw/, */tl/ |
| Coda | Up to 4 consonants, but heavy codas are often simplified in speech | /ksts/ (texts), /lfθs/ (twelfths) | — |
| Onset | /ŋ/ never occurs syllable-initially in English | — | */ŋæt/ is impossible |
Phonotactics explains a great deal of real data. Cluster simplification in connected speech — "texts" /teksts/ reduced to [teks], "asked" reduced to [ɑːst] — is the spoken response to codas that are hard to articulate in full, and it is a connected-speech process rather than an accent feature as such. Phonotactics also shapes how English adapts loanwords: borrowed pronunciations are bent to fit English rules.
/ps/ is illegal in English, so the /p/ is dropped: /saɪˈkɒlədʒi/./ts/ is illegal initially, so the /t/ is dropped: /suːˈnɑːmi/.Not every phonotactically legal string is an actual word. "Blick" /blɪk/ obeys all English rules and could be a word — its absence is an accidental gap. By contrast "bnick" */bnɪk/ violates the rules (/bn/ is not a legal onset) and could never be an English word — a systematic gap. The distinction matters because native speakers have reliable intuitions about it: they recognise "blick" as a possible-but-unused word and "bnick" as impossible, which is direct evidence that phonotactic knowledge is real and internalised.
Key Definition: Phonotactics — the rules governing permissible sequences of phonemes in a language. Accidental gap: a phonotactically legal but non-occurring word (blick). Systematic gap: a string ruled out by phonotactics (bnick).
The heart of this lesson is the set of recurring processes by which underlying forms are realised as surface forms. Each is a regular pattern you can name, evidence and explain in data — and naming the process (not just describing the sound) is exactly what earns AO1 depth.
Assimilation is the process by which a sound becomes more similar to a neighbouring sound, usually to ease articulation. It is overwhelmingly regressive (or anticipatory) in English — a sound is influenced by what follows it — though progressive assimilation (influence from a preceding sound) also occurs.
| Type | What changes | Example |
|---|---|---|
| Place assimilation | A consonant adopts the place of the following consonant | "ten boys" /ten/ → [tem bɔɪz] (/n/ → [m] before bilabial); "good girl" → [ɡʊɡ ɡɜːl] |
| Manner assimilation | A consonant adopts the manner of its neighbour | "in the" → [ɪn nə] in rapid speech |
| Voicing assimilation | A suffix adopts the voicing of the stem-final sound | plural "dogs" /dɒɡz/ (voiced after /ɡ/) vs "cats" /kæts/ (voiceless after /t/) |
The plural and past-tense suffixes are the cleanest examples of systematic voicing assimilation: the regular plural is /z/ after a voiced sound ("dogs"), /s/ after a voiceless sound ("cats") and /ɪz/ after a sibilant ("buses"); the past-tense <-ed> is /d/ after voiced ("played"), /t/ after voiceless ("walked") and /ɪd/ after /t d/ ("wanted"). These are not three random endings but one underlying suffix whose voicing assimilates to the preceding sound — which is why distinctive-feature notation (below) states the rule so economically.
Elision is the deletion of a sound that would be present in a careful citation form. It is pervasive in connected speech and a high-value feature to identify.
| Type | Example |
|---|---|
| Consonant elision (cluster simplification) | "next day" → [neks deɪ] (/t/ elided); "old man" → [əʊl mæn] |
| Schwa elision | "camera" /ˈkæmərə/ → [ˈkæmrə]; "library" → [ˈlaɪbri] |
/h/-elision in weak forms | "tell her" → [tel ə] |
Epenthesis is the insertion of a sound not present underlyingly, again typically to ease articulation between awkward neighbours. A widely cited English example is the intrusive plosive between a nasal and a following fricative: "something" /ˈsʌmθɪŋ/ realised as [ˈsʌmpθɪŋ], or "prince" coming to sound like "prints". The intrusive and linking /r/ of non-rhotic accents ("law-r-and order") is also a kind of epenthesis at the word boundary.
Devoicing is the loss of voicing from a normally voiced sound, most often at the edges of an utterance or next to a voiceless sound. Word-final voiced obstruents are frequently partly devoiced in English: the /d/ of "bed" said in isolation is realised with little or no vocal-fold vibration at its release, approaching [bed̥]. Importantly, English maintains the "voiced/voiceless" contrast in these positions less through actual voicing than through pre-fortis clipping — the vowel before a voiceless consonant is shorter ("bet" has a shorter vowel than "bed"), so vowel length, not consonant voicing, becomes the main perceptual cue.
Key Definition: Assimilation — a sound becomes more like a neighbour. Elision — a sound is deleted. Epenthesis — a sound is inserted. Devoicing — a voiced sound loses its voicing. These connected-speech processes derive surface forms from underlying phonemic representations and are shown in narrow
[ ]transcription.
Subscribe to continue reading
Get full access to this lesson and all 10 lessons in this course.