Have you noticed that many entries in the Morphodex include a morpheme with parentheses followed by two or more forms in brackets? A reader recently asked me what the parentheses and brackets mean. In my post About the Morphodex, I explain, “If a morpheme is allomorphic, the realized morphemes are listed in brackets after the parentheticalized morpheme. For example, the morpheme -(i)l(e) surfaces as the allomorphs -ile, -il, and -le. The entry for the morpheme thus appears as -(i)l(e) [-ile, -il, -le].” But what does any of that explanation mean?
Allomorphs
To understand the parenthetical morpheme, you must first understand the allomorph. The word allomorph consists of the morphemes <All + o + Morph>. The bound base <All> denotes “each other, other, different.” The <o> is a connecting vowel. The free base <Morph> denotes “shape, form.” Thus, allomorph means “different form,” a meaning key to understanding the concept.
Different sources provide slightly different definitions of the allomorph:
- “any of two or more actual representations of a morpheme” (Google)
- “any of the different phonological representations of a morpheme” (Wiktionary)
- “a variant phonetic form of a morpheme, or in other words, a unit of meaning that varies in sound and spelling without changing the meaning” (Wikipedia)
- “one of a set of forms that a morpheme may take in different contexts” (Merriam Webster)
- “each of two or more alternative forms of a morpheme” (Oxford English Dictionary)
- “any of the variant forms of a morpheme as conditioned by position or adjoining sounds” (Collins Dictionary)
- “one of two or more complementary morphs which manifest a morpheme in its different phonological or morphological environments” (SIL Glossary of Linguistic Terms)
A morpheme is the smallest meaningful linguistic unit of a language. An allomorph is one of two or more realizations of a morpheme.
The easiest way to understand allomorphs is to look at the simple past -ed suffix as well as the third person singular simple present suffix and plural noun suffix -(e)s.
Simple Past -ed Suffix
The -ed suffix affixes to the base form of a verb to form the simple past tense (and often the past participle) of a verb. For example:
- <Chat + ed → chatted>
- <Sound + ed → sounded>
- <Kiss + ed → kissed>
- <Wish + ed → wished>
- <Follow + ed → followed>
- <Yell + ed → yelled>
The verbs chat, sound, kiss, wish, follow, and yell all take the -ed suffix to form the simple past tense. But notice that -ed surfaces differently phonologically depending on the phonological context. In other words, speakers pronounce the -ed differently depending on the preceding sound.
In the words chatted and sounded, the -ed surfaces as [әd] because the final sound of chat and sound is a dental [t] or [d]. The -ed surfaces as [әd] to add a syllable to the simple past tense. In English, pronouncing [d] after [t] or [d] is not possible, so the simple past tense of a verb ending in [t] or [d] requires another syllable pronounced as [әd]. The [әd] pronunciation is the first phonological realization of the -ed allomorph.
In the words kissed and wished, the -ed surfaces as the voiceless [t] because the final sound of kiss and wish is a voiceless phone. The voiceless [t] pronunciation is the second phonological realization of the -ed allomorph.
In the words followed and yelled, the -ed surfaces as the voiced [d] because the final sound of follow and yell is a voiced phone. The voiced [d] pronunciation is the third phonological realization of the -ed allomorph.
In the case of the simple past -ed suffix, the <ed> allomorph is realized as [әd], [t], or [d] depending on the phonological environment but spelled the same regardless of the phonological realization. The -ed suffix has one orthographic and three phonological realizations. But sometimes an allomorph surfaces with a different spelling in addition to a different pronunciation.
Third Person Singular Simple Present and Plural Noun -(e)s Suffix
The -(e)s suffix forms the third person singular simple present of verbs and the plural of most nouns. The rules for the -(e)s verb suffix and the -(e)s noun suffix are the same. For example:
- <Box + (e)s → boxes>
- <Fish + (e)s → fishes>
- <Witch + (e)s → witches>
- <Mess + (e)s → messes>
- <Baby + (e)s → babies>
- <Worry + (e)s → worries>
- <Do + (e)s → does>
- <Go + (e)s -> goes>
- <Potato + (e)s → potatoes>
- <Veto + (e)s → vetoes>
- <Bark + (e)s → barks>
- <Bat + (e)s → bats>
- <Jog + (e)s → jobs>
- <Weasel + (e)s -> weasels>
The -(e)s suffix surfaces as the -es allomorph in the words boxes, fishes, witches, and messes. The -es additionally spells [әz] because of the phonological environment. The words box, fish, witch, and mess end in a sibilant ([s], [z], [ʃ], [t͡ʃ], [d͡ʒ]). Sibilants are fricative and affricate consonants produced by the tongue directing a stream of air towards the teeth. The -(e)s suffix cannot surface as -s spelling [s] or [z] after a sibilant because a sibilant cannot follow another sibilant in English. Thus, words that end in a sibilant require another syllable in the third person singular simple present or plural, and the -es allomorph spelling [әz] surfaces.
The -(e)s suffix also surfaces as the -es allomorph after a <y> that toggles with <i> as in babies and worries. The verbs do and go also take the -es allomorph to form the third person singular simple present does and goes. Some verbs and nouns that end in <o> also take the -es allomorph as in potatoes and vetoes. The -es spells the voiced [z] in all three morphological environments.
In all other verbs and nouns, the -(e)s suffix surfaces as the -s allomorph. However, like the -ed suffix, the -s allomorph also surfaces with different pronunciations depending on the final sound of the word to which the suffix affixes. In the words barks and bats, the -s spells the voiceless [s] because the final sound of bark and bat is a voiceless phone. In the words jobs and weasels, the -s spells the voiced [z] because the final sound of job and weasel a voiced phone. The [s] and [z] are the third and fourth phonological realizations of the -(e)s suffix.
The suffix -(e)s has two orthographic realizations (-es and -s) and three phonological realizations ([әz], [s], [z]) depending on the morphological and phonological environment.
Allomorphs and Parenthetical Graphemes
If -s and -es are two different suffixes, why do I represent both with the parenthetical -(e)s suffix? As I have explained, -s and -es are not actually two different suffixes but are instead allomorphs of the same morpheme, the -(e)s suffix. Both forms share the same etymology and meaning. The difference in spelling and pronunciation depends on the environment. The -(e)s suffix is the morpheme; the -s and -es suffixes are the allomorphs.
Could I list -s and -es as separate morphemes? Absolutely! For example:
- <Box + es → boxes>
- <Fish + es → fishes>
- <Witch + es → witches>
- <Mess + es → messes>
- <Baby + es → babies>
- <Worry + es → worries>
- <Do + es → does>
- <Go + es -> goes>
- <Potato + es → potatoes>
- <Veto + es → vetoes>
- <Bark + s → barks>
- <Bat + s → bats>
- <Jog + s → jobs>
- <Weasel + s -> weasels>
However, to maintain consistency across the Morphodex, I chose to represent allomorphic variants with parentheses whenever possible. I also use parenthetical forms to maintain relationships between morphemes and evidential words.
While the relationship between -s and -es is generally clear and understood, the same is not true of all morphemes. Take the -(i)l(e) suffix as an example. The -(i)l(e) entered English via three closely-related etymological routes: (1) Old French -il or Latin -īlis; (2) French -ile and German -il from Latin -īlis; (3) French -ile from Italian -ile from Latin -īle from -īlis. The ultimate source of English -(i)l(e) is Latin -īlis, but different words with the same suffix entered English in different ways.
The -(i)l(e) suffix surfaces as -ile as in servile, -le as in noble, and -il as in civil. The form of the allomorph depends on the specific word. For example, servile comes from Middle English servyle from Old French servile from Latin servīlis. Noble comes from Middle English noble from Old French noble from Latin nōbilis. Civil comes from Middle English civil from Old French civil from Latin cīvīlis. The realization depends on the etymology of the word.
- <Serve + (i)l(e) → servile>
- <Nobe + (i)l(e) → noble>
- <Cive + (i)l(e) → civil>
Furthermore, the words are servile, noble, and civil, not *servil or *servle, *nobile or *nobil, or *civile or *civle. The -(i)l(e) surfaces as -ile to spell servile, -le to spell noble, and -il to spell civil. The parenthetical -(i)l(e) maintains the etymological and allomorphic relationship between -ile, -le, and -il. Could I create separate entries for -ile, -le, and -il? Again, absolutely. But the parenthetical -(i)l(e) maintains the relationship between the three allomorphs.
Additional Allomorph Examples
In addition to affixes, bases can also surface as two or more allomorphs. Take the bound base <Arch(a)e> as an example. Denoting “ancient, primitive, old, related to the beginning,” <Arch(a)e> comes from the Latinized form of Ancient Greek arkhaîa from arkhaîos from arkhḗ from árkhō. Historically, the base was spelled <Archæ> with the <æ> as the final vowel. The realization <Archae> maintains the historical spelling. British English more commonly maintains the <Archae> allomorph as in archaeology. American English prefers the more modern <Arche> realization with the pronounced final <e> as in archeology. The parenthetical <Arch(a)e> is the bound base of the allomorphs <Archae> and <Arche>.
- <Arch(a)e + o + Log(u)(e) + y → archaeology>
- <Arch(a)e + o + Log(u)(e) + y → archeology>
Another example is the bound base <Mon(e)>, which surfaces as the allomorphs <Mone> and <Mon>. Denoting “one, alone, only, sole, single,” <Mon(e)> comes from Ancient Greek mónos. The replaceable <e> surfaces when needed in <Mone> and does not surface when not needed in <Mon>. For example:
- <Mon(e) + ad → monad>
- <Mon(e) + epi + Scope + acy → monepiscopacy>
- <Mon(e) + Arch → monarch>
- <Mon(e) + o + Clone + al → monoclonal>
In the word monad, the <Mon(e)> surfaces as <Mone> with the replaceable <e> to prevent the <n> from doubling. Without the replaceable <e>, the vowel-initial -ad suffix would cause the <n> to double.
- <Mone + ad → monad>
- <Mon + ad → *monnad>
In the words monepiscopacy and monarch, the <Mon(e)> surfaces as <Mon> without the replaceable <e>. Prefixes and bases do not replace a replaceable <e>. If the <Mon(e)> surfaced as <Mone>, the <e> would erroneously remain in the final word.
- <Mon + epi + Scope + acy → monepiscopacy>
- <Mone + epi + Scope + acy → *moneepiscopacy>
- <Mon + Arch → monarch>
- <Mone + Arch → *monearch>
The <Mon(e)> can surface as either <Mon> or <Mone> in monoclonal because a connecting vowel replaces a replaceable <e> but does not cause doubling.
- <Mon + o + Clone + al → monoclonal>
- <Mone + o + Clone + al → monoclonal>
Again, I could create separate entries for <Mon> and <Mone>, but then I would lose the relationship between words in which the <Mon(e)> surfaces as <Mon> like monepiscopacy and monarch and words in which the <Mon(e) surfaces as <Mone> like monad. And would words like monoclonal with a connecting vowel after the base be evidence of <Mon> or <Mone>? The parenthetical <Mon(e)> maintains the allomorphic relationship between <Mon> and <Mone>.
Thus, to answer the question about the parentheses and brackets in Morphodex entries, the parentheses indicate that a morpheme surfaces as two or more allomorphs. The brackets show the possible allomorphic realizations of the morpheme.
