Skip to main content
Back to path

Welcome to a Grasp lesson

Create a free account to keep studying and create your own study path.

Lesson 6

Allophones or Distinct Phonemes?

Recommended resources

Hello! Welcome to your sixth lesson.

In our last session, you learned the crucial difference between phonemes (abstract sound categories) and allophones (their physical pronunciations). We established a systematic procedure for analyzing phonetic data, using the principles of contrastive and complementary distribution to determine the relationship between sounds.

Today's lesson builds directly on that foundation. It is designed to meet the learning outcome: Analyze phonemic data to determine if two sounds are allophones of the same phoneme or distinct phonemes. We will move from introductory examples to more complex datasets, honing your analytical skills. This is less about learning new concepts and more about gaining fluency in the analytical process itself—a skill that is fundamental to historical linguistics.

By the end of this lesson, you will be more confident in tackling raw linguistic data, identifying underlying patterns, and making principled arguments about a language's sound structure.

1. Recap and a More Refined Toolkit

Let's quickly review the core logic from our last lesson:

  • Minimal Pairs → Contrastive Distribution → Separate Phonemes. If two sounds can appear in the exact same environment to create two different words (e.g., English pat vs. bat), they are distinct phonemes.
  • No Overlap → Complementary Distribution → Allophones of One Phoneme. If two phonetically similar sounds never appear in the same environment, they are allophones of a single phoneme (e.g., English aspirated [tʰ] and unaspirated [t]).

But what happens when the data is messy? What if you can't find a perfect minimal pair, but the sounds still seem to be used unpredictably? For this, we need to add a concept to our toolkit: the near-minimal pair.

2 Phonemic Analysis

The following reading from the 'Phonemic Analysis' PDF introduces the idea of near-minimal pairs. This is a useful tool when perfect minimal pairs are absent.

Please read the section '2.10.2 Near-minimal pairs'. Focus on the logic: if two sounds appear in very similar environments and no predictable rule can be found to explain their distribution, they are likely separate phonemes, even without a perfect minimal pair.

A near-minimal pair shows that two sounds occur in such similar contexts that a rule for complementary distribution would be impossible to formulate. For example, in the reading's case of English /ð/ (as in thy) and /ʒ/ (as in measure), even if no perfect minimal pair exists for a speaker, words like leather [lɛðɚ] and leisure [liʒɚ] show that both sounds can occur between vowels. This makes a complementary distribution analysis highly unlikely.

So, our analytical possibilities are:

  1. Contrastive Distribution: Sounds distinguish meaning (minimal/near-minimal pairs). → Separate Phonemes.
  2. Complementary Distribution: Sounds are in mutually exclusive environments. → Allophones of one Phoneme.
  3. Free Variation: Sounds can be swapped in the same environment with no change in meaning. This is less common and often involves very subtle phonetic differences. We will treat it as a minor possibility for now.

2. Guided Analysis: A Deeper Dive with Maasai

Now, let's apply our systematic procedure to a more extensive dataset from Maasai, a language spoken in Kenya and Tanzania. The process of organizing this data and uncovering the underlying system is an excellent exercise in phonological problem-solving.

We will start by focusing on the velar sounds in the data: the voiceless stop [k], the voiced stop [g], and the voiced fricative [ɣ] (represented in the resource by [“]).

First, let's get familiar with the data and the method for organizing it.

2 Phonemic Analysis

This section of the 'Phonemic Analysis' PDF presents the Maasai data and demonstrates the method of compiling 'local environments' for each sound of interest. This is a more formal application of the environment-listing technique we used in the last lesson.

Please read the section '2.10.3 Using local environments to establish complementary distribution'. Read from the beginning of the section down to the table that lists the environments for [k], [g], and [“]. Your goal is to understand how the table of environments was constructed from the word list.

Now that you've seen how the environments are listed, let's analyze them together.

Step 1: Examine the environments for [k], [g], and [ɣ]

Here is the table of environments from the reading:

[k] [g] [ɣ] ([“])
# __ a ŋ __ a a __ e
# __ e ŋ __ i a __ i
# __ i ŋ __ o i __ a
# __ o i __ i
l __ e i __ o
l __ i i __ u
r __ e o __ i
r __ i o __ op
a __ # u __ a
ep __ #
o __ #
op __ #

(Note: [“] is used for [ɣ], # for word boundary)

Step 2: Check for Contrastive Distribution

Do any of these sounds appear in the same environment?

  • A quick scan shows no identical environments. For example, [g] only ever appears after [ŋ], while [k] and [ɣ] never do. [ɣ] only appears between vowels, but [k] can appear at the beginning or end of a word. There are no minimal or near-minimal pairs.

Step 3: Look for Complementary Distribution

Since the environments don't overlap, let's define them.

  • Environment for [g]: The pattern is perfectly consistent. [g] occurs only after the velar nasal [ŋ].
  • Environment for [ɣ]: This sound appears in environments like a __ e, i __ a, o __ i. The pattern is that [ɣ] occurs only between two vowels.
  • Environment for [k]: This sound appears everywhere else: at the beginning of a word (# __ a), at the end of a word (a __ #), and after consonants like [l] and [r].

The three sounds are in complementary distribution. Their environments are mutually exclusive.

Step 4: State the Conclusion and Formulate the Rules

Since [k], [g], and [ɣ] are phonetically similar (they are all velar consonants) and are in complementary distribution, we conclude they are allophones of a single phoneme.

The "elsewhere" case, [k], is the most general and is considered the default allophone. We use its symbol to represent the phoneme: /k/.

The analysis can be summarized with these rules:

  1. /k/ → [g] / [ŋ]__ (The phoneme /k/ is pronounced as [g] after [ŋ])
  2. /k/ → [ɣ] / V__V (The phoneme /k/ is pronounced as [ɣ] between vowels)
  3. /k/ → [k] elsewhere.

3. Practice and Generalization

Now it's your turn. The Maasai data also contains bilabial sounds ([p], [b], [β]) and dental sounds ([t], [d], [ð]).

Your Task:
Using the same Maasai word list, analyze the distribution of the bilabial sounds [p], [b], [β] and the dental sounds [t], [d], [ð].

  1. Are they allophones of single phonemes? Or are they distinct?
  2. If they are allophones, what are the phonemes and what are the rules for their distribution?
  3. What do you notice when you compare the pattern for the bilabials, dentals, and the velars we just analyzed?

Take about 10-15 minutes to work through this. You don't need to write out every single environment, but try to spot the general pattern for each sound.

Click to see the analysis

If you examine the data, you will find a strikingly parallel pattern.

Bilabials: [p], [b], [β]

  • [b] occurs only after the bilabial nasal [m]. (e.g., [embiβan])
  • [β] occurs only between vowels. (e.g., [enaiβoàa])
  • [p] occurs elsewhere (word-initially, word-finally, etc.). (e.g., [ailap])
  • Conclusion: [p], [b], [β] are allophones of the phoneme /p/.

Dentals: [t], [d], [ð]

  • [d] occurs only after the dental/alveolar nasal [n]. (e.g., [indai])
  • [ð] occurs only between vowels. (e.g., [ema¢a¨a])
  • [t] occurs elsewhere. (e.g., [aret])
  • Conclusion: [t], [d], [ð] are allophones of the phoneme /t/.

The Generalization

The most important insight here is that the same processes are happening to all three sets of sounds. This is not a coincidence. The sounds [p, t, k] form a natural class: the voiceless stops. The sounds [b, d, g] form the natural class of voiced stops, and [β, ð, ɣ] are voiced fricatives.

Instead of writing three separate sets of rules, we can write two general rules that apply to the entire natural class of voiceless stops. This is a much more powerful and elegant explanation, a principle you're likely familiar with from mathematics and programming.

2 Phonemic Analysis

This final section on the Maasai data explains how to generalize the analysis using natural classes and phonetic features.

Please read the section '2.10.4 More Maasai: natural classes'. It shows how the patterns you just found can be combined into general rules, a core principle of phonology.

The general rules for Maasai are:

  1. Postnasal Voicing: A voiceless stop becomes voiced after a nasal consonant.
  2. Spirantization: A voiceless stop becomes a voiced fricative between vowels.

4. A Cross-Linguistic Perspective: Same Sounds, Different Systems

To prepare for comparing languages historically, it's vital to understand that languages can use the same phonetic sounds but organize them differently. A distinction that is allophonic in one language can be phonemic in another.

2 Phonemic Analysis

This reading compares the status of sounds like [t], [d], [ɾ], and [ð] in English and Spanish. It's a classic example of how phonological systems can differ.

Please read the section '2.9 Phonemes in Other Languages', focusing on the comparison between English and Spanish. Pay attention to the diagrams at the end that visually represent how the two languages 'group' the same phonetic space differently.

This comparison reveals a critical point for historical linguistics:

  • In English, [t] and [ɾ] are allophones of /t/. But in Spanish, they are separate phonemes, /t/ and /ɾ/, as shown by the minimal pair pita vs. pira.
  • In English, [d] and [ð] are separate phonemes, /d/ and /ð/ (die vs. thy). But in Spanish, they are allophones of /d/.

When we compare related languages, we will find cases where one language preserves a distinction (e.g., two separate phonemes) that another language has lost (by merging them into one phoneme) or turned into an allophonic variation. Identifying these differences is how we begin to reconstruct their shared history.

Conclusion

In this lesson, we have moved beyond introduction and into application, strengthening your ability to perform phonemic analysis.

Key Takeaways:

  • Phonemic analysis is a systematic process of listing environments to check for contrastive or complementary distribution.
  • Near-minimal pairs can be used as evidence for distinct phonemes when perfect minimal pairs are unavailable.
  • Phonological rules often apply to natural classes of sounds (e.g., all voiceless stops), allowing for powerful generalizations.
  • Languages can have similar phonetic inventories but organize them into different phonological systems, a key source of variation and historical change.

You are now well-prepared to analyze the sound systems of languages from a synchronic (single point in time) perspective.

Preview of the Next Lesson:
In our next lesson, we will begin our journey into diachronic (historical) analysis. The learning outcome will be to "Classify sound changes as unconditioned, conditioned, assimilation, dissimilation, or deletion." We will see how synchronic allophonic rules, like the ones we formulated today, often provide a snapshot of a historical sound change in progress.