Table of Contents
Phonology and spelling
An auxiliary language should have a phonology that’s fairly average – it shouldn’t have more sounds that the average language (though it may have less) and it should only have the vowels and consonants that are most common among the world’s language, arranged in syllables that aren’t more complex than what’s average among the world’s languages.
Its spelling should use the globally most widespread writing system (the Latin script) and the spellings used for each sound should be easy to recognize for a large number of people as well as easy to type.
Lugamun’s phonology and spelling were developed on these principles. They are based on information found in WALS and PHOIBLE, a repository of the phonemes (sounds) that can be found in the world’s languages.
Vowels and diphthongs
According to WALS, the average number of vowels used by the world’s languages is slightly below six (WALS 2 – read: WALS, chapter 2). If we round this down, it means that our language should have no more than five vowels – which is also by far the most frequent size of the vowel inventory among the world’s languages (ibid.). We allow the five vowels that occur in at least 60 percent of the world’s languages, according to PHOIBLE:
- a /a/ as in Spanish or Italian casa, and like or similar to the a in ‘father’ and for many (especially British) speakers in ‘bat’ (open central or front unrounded vowel).
- e /e/ as Spanish bebé, French fée, or the e in ‘hey’ – but without the following i-like sound (mid or close-mid front unrounded vowel).
- i /i/ as in ‘free’ or Spanish tipo (close front unrounded vowel).
- o /o/ as in Spanish como or French sot, and like or similar to the o in ‘tore’ (mid or close-mid back rounded vowel).
- u /u/ as in ‘boot’ or Spanish una (close back rounded vowel).
The vowels may be considered as arranged in the following chart:
- No other vowel occurs in more than 37 percent of the world’s languages, making this a very clear choice.
- This vowel system corresponds to several typical features as described by WALS: There are no contrastive nasal vowels and no front rounded vowels (WALS 10–11). Tone is not a distinctive feature of words (WALS 13).
- Though derived independently, this vowel system also corresponds well to the phonetics of typical creole languages as analyzed by APiCS, the Atlas of Pidgin and Creole Language Structures: there are no tone distinctions (APiCS 120), no nasal vowels (122), and no schwa (123).
Diphthongs are two vowels that are pronounced jointly as part of the single syllable. The first vowel is pronounced as usual, followed immediately by the second vowel, which is pronounced quickly and without stress. Neither WALS nor PHOIBLE has clear information on diphthongs, but another database called LAPSyD does. Following this database, we accept three diphthongs into our phonology:
- ai /ai̯/ – similar to the vowel in ‘price’
- au /au̯/ – similar to ‘mouth’
- oi /oi̯/ – similar to ‘choice’
In cases where a combination of vowels looks like one of these diphthongs, but should actually be read as two separate vowels that belong to different syllables, an apostrophe is inserted between the two letters to make the intended pronunciation clear: o’i represents two syllables, while oi represents just one.
- To see diphthong frequencies, follow the LAPSyD link given above, then select “Aggregate Vowel inventory” instead of “Show Language list” and click “show visualization”. To sort the results, click on the “count” column in the “Diphthongs” table. Four diphthongs occur in at least twelve of the investigated languages. The rarest of these – /ei̯/ – is formed of vowels that are directly next to each other in the vowel chart given above. In the case of such related vowels the risk is higher that people will clearly articulate just one half of the diphthong (reducing /ei̯/ to /e/), therefore we don’t admit this diphthong, but we accept the other three.
- The use of the apostrophe as a vowel separator is inspired by pinyin.
- Some linguists distinguish between “falling diphthongs” – as described here – and “rising diphthongs” which are sequences of an approximant (or semivowel) followed by a vowel. The latter will be covered below.
According to WALS, the median number of consonants among the world’s language is 21, while the average (rounded down) is 22 (WALS 1). We should admit no more than that to keep our language fairly easy to pronounce for most people. We allow most of the consonants that occur in at least 25 percent of the world’s languages, according to PHOIBLE – with some restrictions motivated below. This results in a core set of 18 consonants:
- b /b/ as in ‘bus’ (voiced bilabial plosive).
- c /t̠ʃ/ as in ‘child’ (voiceless palato-alveolar sibilant affricate).
- d /d/ as in ‘dog’ (voiced alveolar or dental plosive).
- f /f/ as in ‘fish’ (voiceless labiodental fricative).
- g /g/ as in ‘get’ (voiced velar plosive).
- h /h/ as in ‘high’ (voiceless glottal fricative). May also be pronounced /x/ as in Scottish English ‘loch’ or German Buch (voiceless velar fricative).
- j /d̠ʒ/ as in ‘jump’ (voiced palato-alveolar sibilant affricate).
- k /k/ as in ‘kiss’ (voiceless velar plosive).
- l /l/ as in ‘leg’ (alveolar or dental lateral approximant).
- m /m/ as in ‘mad’ (bilabial nasal).
- n /n/ as in ‘nine’ (alveolar or dental nasal).
- p /p/ as in ‘pick’ (voiceless bilabial plosive).
- r /ɾ/ as in Spanish caro (voiced alveolar tap or flap). May also be pronounced /r/ as in Spanish perro (voiced alveolar or postalveolar trill, “rolled R”) or /ɹ–ɻ/ as in English ‘red’ (voiced alveolar, postalveolar, or retroflex approximant).
- s /s/ as in ‘sit’ (voiceless alveolar sibilant). May also be pronounced /z/ as in ‘zoo’ (voiced alveolar sibilant).
- t /t/ as in ‘tape’ (voiceless alveolar or dental plosive).
- v /w/ as in ‘weep’ (voiced labio-velar approximant). May also be pronounced /v/ as in ‘view’ (voiced labiodental fricative).
- x /ʃ/ as in ‘sheep’ (voiceless palato-alveolar sibilant).
- y /j/ as in ‘you’ (voiced palatal approximant).
The voiceless plosives (k, p, t) may be pronounced with aspiration, as frequently used in certain English words such as ‘pin’, and as in Chinese 口 (kǒu), 旁 (páng), 透 (tòu). The absence or presence of aspiration does not signal a difference in meaning.
Four other consonants are optional:
- Adjacent vowels that don’t form a diphthong should be pronounced clearly separate from each other, as they belong to different syllables. Optionally a glottal stop, /ʔ/ – as in the middle of ‘uh-oh’ – may be pronounced between such vowels. Either pronunciation is fine, and if you don’t know what a glottal stop is, don’t worry about it.
- Before g or k, n is commonly pronounced as /ŋ/ as in ‘sing’ (velar nasal) rather than /n/ – e.g. inglis /iŋˈglis/ ‘English’. The usual pronunciation /n/ is also fine, however.
- The combination ny may be pronounced as /nj/ – the sequence of the two consonants which these two letters usually represent – or as the single consonant /ɲ/, as in Spanish enseñar or Swahili nyama (voiced palatal nasal). Either pronunciation is fine.
- The combination ts may be pronounced as /ts/ – the sequence of the two consonants which these two letters usually represent – or as the single consonant /t͡s/ (voiceless alveolar sibilant affricate), as in Mandarin 早 (zǎo) or Russian царь (carʹ). Either pronunciation is fine. For the purposes of hyphenation, t is still considered to end a syllable and s to start the next one, even if both are pronounced as a single sound.
The letters q, w and z are not used, except in proper names and foreign words.
The following chart shows all the non-optional consonants:
|Plosive||p b||t d||k g|
|Affricate||t̠ʃ (c) d̠ʒ (j)|
|Approximant||l||j (y)||w (v)|
If the spelling differs from the IPA symbol, it is shown in parenthesis.
- /z/ and /v/ occur in 27–30% of the languages listed in PHOIBLE. But they are rarer than their voiceless equivalents /s/ and /f/, and a voicing contrast exists most typically in plosives, but not in fricatives, which include the sibilants (WALS 4). So, to avoid introducing such a voicing contrast, we don’t admit these sounds as separate phonemes. /z/ is admitted as a variant pronunciation of its voiceless equivalent.
- /v/ could conceivably be considered an acceptable alternative to several other consonants. Its voiceless equivalent /f/ would be one candidate, while speakers of languages exposing the widespread phenomenon known as betacism might consider it most similar to /b/, and speakers of languages that treat /v/ and /w/ as allophones – such as Hindustani – might consider it most similar to /w/. Historically, the letter w has developed out of v. In the Classical Latin alphabet, v represented both the vowel /u/ and the semivowel /w/. Only in later times did the pronunciation of that semivowel shift to the fricative /β/ (still widespread in modern Spanish as an allophone of /b/) and in other Romance languages further to /v/. Because of the Hindustani allophony of /v/ and /w/, the historical development of /v/ in modern Romance (and related) languages out of classical /w/, and the visual similarity and joint origin of the letters v and w, we treat /v/ as a acceptable alternative of /w/. Accordingly, /v/ in the source languages typically becomes /w/ in Lugamun (written v for reasons that will be further explained below).
- All additional sounds occurring in at 18 percent of the world’s languages are admitted as alternative pronunciations of the sound or sound combination to which they can be considered most similar.
- As rhotic consonants vary a lot between the world’s languages and it might be hard for people to get used to new ones, we allow the three most common rhotic consonants as pronunciations of r. But which one should be the preferred pronunciation? According to the PHOIBLE data, the trill /r/ might be the most widespread pronunciation (44%), followed by the tap or flap /ɾ/ (26%). However, /r/ is probably overcounted at the cost of other rhotics because of transcribers sometimes using the simple letter /r/ instead of the IPA letters for other rhotics, which are less easy to type. Among our source languages, a tap or flap is actually somewhat more common than the trill. Japanese, Spanish, Swahili, and some widespread Arabic variety such as Egyptian and Moroccan Arabic use the alveolar tap or flap /ɾ/, while Hindustani uses the voiced retroflex flap /ɽ/. On the other hand, Standard Arabic (but not all dialects), Indonesian, Russian, and Spanish use the alveolar or postalveolar trill /r/. While this is a close call, the tap or flap seems a bit more common cross-lingustically; moreover, it is arguably easier to learn for those not used to either than the trill, and more similar to the approximant /ɹ–ɻ/, which is used in English and (by some speakers) in Mandarin Chinese. Therefore we recommend the tap or flap as standard pronunciation, while admitting the other variants as alternatives.
- Aspired plosives are relatively rare – they occur only in 20 percent or less of the world’s languages – therefore they are only allowed as alternative pronunciations.
- /ʔ/ and /ɲ/ are kept optional to avoid difficult-to-distinguish “minimal pairs” – words that differ only in the absence or presence of a glottal stop between vowels or in the usage of /nj/ versus /ɲ/.
- The velar nasal /ŋ/ is considered an alternative pronunciation of /n/ in certain positions rather than an independent phoneme because only a minority of languages allow it at the start of syllables (WALS 9) and our rules for syllable structure – as explained below – don’t allow it at the end of syllables either. Hence no position remains where it could occur as an independent phoneme. Instead we allow the alveolar or dental nasal /n/ to become velar before the velar plosives – a combination that is common in English, Hindi, Indonesian, Mandarin, and other languages.
- Without requiring further changes, our consonant inventory corresponds to several other features analyzed as most typical by WALS. There are six plosives: /p, t, k, b, d, g/ (WALS 5). The only lateral consonant is /l/ (WALS 8). There are no uvular consonants and no glottalized consonants (WALS 6–7). There are no clicks, labial-velars, pharyngeals, or ‘th’ sounds (WALS 19).
To avoid confusion between words that are pronounced, the core vocabulary will generally not include pairs of words that differ only in small details. Specifically, this means:
- Pairs where one word has the voiceless affricate c and the other has its voiced equivalent j. Since we have can ‘give birth’, we won’t have jan; and since we have jenis ‘gender’, we won’t have cenis.
- Pairs where one word has a vowel, while the other has the equivalent semivowel. Since we have ya ‘he, she’, we won’t have ia.
- Pairs that differ only in the presence or absence of an apostrophe.
Notes on the spellings:
- The above spellings are based on three criteria: avoid diacritics to be easy to type for everyone (many Latin-based languages use some diacritics, but they generally don’t agree on which ones); follow the “one sound – one letter” principle where it is reasonable to do so; and use representations that are already well-known from widely spoken languages. The vowel spellings are obvious, as the five vowel sounds correspond to the five vowel letters in the Latin alphabet in a self-evident way. Most consonant spellings are also quite obvious – in all cases where English and the International Phonetic Alphabet (IPA) agree on a spelling, other Latin-based languages tend to use the same spelling, which can therefore be used without requiring further discussion. The five consonants where this it not the case will be discussed next. In these cases, the resolution is to use one of the spellings that are most common among the most widely spoken languages using the Latin alphabet, but preferring single letters over sequences of two (or more) letters if both are used. The following analysis is based on those of the 25 most widely spoken languages that use the Latin alphabet (English, French, German, Hausa, Indonesian/Malay, Javanese, Portuguese, Spanish, Swahili, Turkish, Vietnamese). Additionally pinyin, the romanization of the most widely spoken language that uses another writing system, is considered as well. Since only five of our ten source languages use the Latin alphabet (English, French, Indonesian, Spanish, Swahili), we need to consider a wider basis here.
- /t̠ʃ/ is written c in Hausa, Indonesian, and Javanese. English, Spanish, and Swahili use ch, but we prefer the representation that uses just one letter.
- /d̠ʒ/ is written j in English, Hausa, Indonesian, Javanese, and Swahili. No two other considered languages share the same common representation, making this the obvious choice.
- /k/ is written k in German, Indonesian, Javanese, pinyin, Swahili, and Turkish. In English and Vietnamese, it is usually c or k, depending on context (the sound that follows); in French, Portuguese, and Spanish it is usually c or qu, depending on context. c might be considered an alternative, but those languages that use c for /k/ use that spelling only in certain contexts, while c before front vowels such as e and i is typically pronounced /s/ or similar. This would make misreadings likely if c were used everywhere. qu would be a conceivable alternative, but it is much less common than k and uses one letter more without any obvious advantage.
- /w/ is written w in English, Hausa, Indonesian, Javanese, and Swahili; w or u (after initials) in pinyin; u in Portuguese and Spanish (typically after /k/ or /g/ or in diphthongs); u or sometimes o in Vietnamese (in the same positions); ou or sometimes w in French. The most common spelling would therefore be w, closely followed by u (which, however, is needed for the vowel). But the allowed alternative pronunciation of /w/ is /v/ (most typically written v in languages using the Latin alphabet), and it would be somewhat odd to have w, but no v, considering that w is graphically “double v”, so it takes a bit more space in printing and a bit more time in handwriting than a single v would. Historically, the letters u, v, and w all have developed out of v, which used to represent both /u/ and /w/ (but not /v/) in the Classical Latin pronunciation (as noted above). To avoid the “pointless” doubling that having just w would entail, and as a nod to the Classical Latin alphabet, we therefore write /w/ as v. This can also be considered a “compromise” between w (double v) and u (visually similar to v and of the same origin, but needed for the vowel), these being the two most common representations of this sound. Finally, writing v facilitates the integration of words from European languages that have /v/ (such as eviden ‘obvious, evident’); while this sound is converted to /w/, the spelling is typically preserved.
- /ʃ/ is written as sh in English, Hausa, and Swahili; as ch in French; as ch or x in Portuguese. x is also used in several other Romance languages. Standard Chinese doesn’t have /ʃ/, but pinyin uses both x and sh for quite similar sounds – x for /ɕ/, the voiceless alveolo-palatal sibilant fricative, and sh for /ʂ/, the voiceless retroflex sibilant fricative. We prefer the single letter x over the digraphs.
- /j/ is y in English, Hausa, Indonesian, Javanese, Swahili, Turkish, and occasionally also in French, Portuguese, Spanish, and Vietnamese. No two other considered languages share the same common representation, making this the obvious choice.
Syllable structure and hyphenation
According to WALS the most typical and median syllable structure among the world’s languages may be called “moderately complex” (WALS 12). Except for proper names, all words in our language should correspond to this structure. This means that syllables may have the form (C)V(C), where C represents a consonant and V a vowel (which might be a diphthong). In other words, syllables consist in a vowel which is optionally followed and/or preceded by a consonant.
The form CCV(C) is also allowed, but only if the second consonant is a liquid (l or r) or a semivowel (v or y). Lugamun’s phonology is further restricted by only allowing the following pairs of consonants at the start of syllables:
- bl, fl, gl, kl, pl, sl
- br, dr, fr, gr, kr, pr, tr
- by, cy, fy, ky, my, ny, py, xy
- cv, dv, gv, hv, kv, sv, tv, xv
Note that v and y can be considered as consonantal equivalents of the vowels u and i. If you don’t know how to pronounce them or have difficulties pronouncing them in any of these clusters, just pronounce the vowel quickly and without stress, followed by the actual vowel which forms the core of the syllable.
All syllables end in either a vowel or one of the consonants l, m, n, r, s, t. Sequences or two or more consonants don’t occur at the end of syllables. If you find it difficult to pronounce any of the allowed consonants in a syllable-final position or to pronounce a cluster of three consonants that might result if a syllable ending in a consonant is followed by one that starts with two, you might add an unstressed neutral vowel (the so-called schwa /ə/, as at the start of ‘about’) or e at the end of the syllable.
Though they would be allowed by the rules listed above, the consonant combinations ry, sy, ty are avoided in Lugamun. Instead the semivowel is replaced by the corresponding vowel i in such cases (ri, si, ti), for example in nasion ‘nation’ and sosieti ‘social’. In these and other cases you may pronounce an unstressed i or u followed by another vowel as the corresponding semivowel (y or v) if you wish. Hence nasion may be pronounced as /nasiˈon/ or as /nasˈjon/, just as you prefer.
- The specific set of consonants allowed to end a syllable was chosen on the basis of our source languages. Only consonants that commonly occur in a word-final position in at least half of them were accepted, with the further requirement that at least two of the source language that allow them must be non-Indo-European. The latter restriction was motivated by the fact that Indo-European languages tend to be much more generous in the set of final consonants they accept than other languages, at least among our sources. As Japanese, Mandarin, and Swahili are particularly restrictive regarding final consonants, the practical result is that the final consonants that commonly occur in both Arabic and Indonesian are allowed in our phonology as well.
- There is only one consonant allowed in two or more non-Indo-European source languages that fails the “half of all our source languages” criterion: the velar nasal /ŋ/, which in a word-final position can only be found in English, Hindi, Indonesian, and Mandarin. Since it is also rare at the beginning of syllables (WALS 9), this means that no position remains where it could occur as an independent sound. Hence we only accept it as an optional consonant into our phonology.
- The consonant pairs allowed to start syllables are those that occur in this position (more frequently than as rare exceptions) in at least five of our ten source languages. Moreover, consonant pairs that occur in this position in Mandarin Chinese are also allowed even if they only occur in two or three other source languages. This adds adds six clusters ending in one of the semivowels -v and -y that would otherwise not be allowed (cv, cy, hv, tv, xv, xy). The reason for these additional admissions is that such consonant–semivowel pairs are very widespread in the Chinese vocabulary, where each core concept tends to be represented by a single syllable. Changing the semivowel to a vowel in such cases (hence dividing the single syllable into two) would make words of Chinese origin much less recognizable.
- Two additional pairs that would fulfill the above criteria have, however, been excluded: dy because in rapid speech it can sound quit similar to j, and ty because it can sound similar to c. For the same reason, ty is also avoided between vowels, where it could otherwise still occur (since t is allowed to end a syllable). Likewise, sy is avoided between vowels because in rapid speech it can sound quit similar to x. The combination ry is avoided since it could be quite hard to pronounce, especially if one speaks the r as an approximant, as usual in English and Mandarin.
As in all languages using the Latin alphabet, words can be divided at syllable boundaries to better fill the line. If syllables are separated by an apostrophe, the word is simply broken after the apostrophe; otherwise a hyphen is added before the line break. The following rules are recommended for hyphenation:
- Never separate the first or the last letter from the rest of the word. For example, don’t divide avan (its syllable structure is a-van); divide beria as be-ria, but not as beri-a.
- If there is just one consonant between vowels, it belongs to the second syllable, e.g. du-bu, glu-hoi, na-ya.
- If there are two consonants between vowels and the first of them is allowed at the end of syllables, divide between them, e.g. ar-bol, las-te, sen-to.
- Otherwise the last two consonants preceding a vowel belong to the second syllable, e.g. sa-fra, in-glis.
- These rules apply to words that have been incorporated into Lugamun’s phonology and spelling. Non-adapted words from other languages should be hyphenated according to the syllable structure of the language they come from instead, e.g. Wash-ing-ton, Frank-furt.
If a word has two or more syllables, one of its last three syllables carries the main stress. The stressed vowel is always the last vowel before the last consonant. If there is no such vowel, the first vowel is stressed.
Examples where the last syllable is stressed (the stressed vowel is always printed in bold):
aual, dodes, duan, hitam
Examples where the next-to-last syllable is stressed:
akai, busu, dulse, kula, situasi
Examples where the third last syllable is stressed:
Note: “No fixed stress” is the most common general stress rule, considerably more frequent than any placement of stress on the same syllable in all words (WALS 14). Among languages who don’t have fixed stress, a relative majority has “stress on ultimate or penultimate syllable” (WALS 15). We somewhat deviate from this pattern, by allowing the third last syllable to be occasionally stressed as well (WALS calls this “right-oriented”), since the resulting rule is easy to describe and remember and fits our vocabulary quite well.
In long words, every syllable that separated by an odd number of other syllables from the stressed one may be considered as carrying secondary (less strong) stress (WALS 17). In the following example, the vowel carrying secondary stress is printed in italics:
Generally, word stress (also called “lexical stress”) is much less important than in English, and if you stress all syllables evenly, that’s quite acceptable as well. However, if you stress a syllable, it should be the one described here.