| Voice onset time |
|---|
| + Aspirated |
| 0 Tenuis |
| − Voiced |
The terms fortis and lenis (in the wide sense) apply to this opposition regardless of whether it's only an opposition of voice or not. Using English as an example, this means that when we devoice a voiced consonant, (such as 'd' at the end of a word,) for ease of articulation, it still retains the fortis / lenis distinction. This is not true of Polish, where the voiced consonant retains only the place of articulation and becomes fortis. Thus the terms fortis and lenis allow one to describe in more precise terms than 'voiced and unvoiced' the articulation of French, English, Polish and southern German consonants. They refer to a bundle of articulatory features which have different distributions in different languages. Not all of them need to be present in a particular language:
It is commonly said that tenseness is what distinguishes fortis and lenis in the narrow sense: In the articulation of the fortis, more muscular energy is supposedly used. However, this has never been demonstrated.
In Korean, a higher fundamental frequency of vowels following certain 'tense' consonants is thought to be a result of increased muscular tension in the vocal cords, a phonation called stiff voice. However, in Swiss German, no possible acoustical correspondent of the assumed tenseness has been found. Consequently, it is debated whether the Swiss German opposition is really based on different muscular tension, and not on gemination.
The North Caucasian languages (Circassian and Dagestanian) have a consonantal distinction described as strong or preruptive that has concomitant length. Akhvakh and other Dagestanian languages even possess a distinction between strong/long and weak/short ejective consonants: [qʼaː soup, broth vs. [qːʼama cock's comb. (Tense phonemes in these languages are traditionally transcribed with the length diacritic, following the Cyrillic orthography of these languages.) Kodzasov (1977:228, translated in L&M 1996:97–98) describes them for Archi: "Strong phonemes are characterized by the intensiveness (tension) of the articulation. The intensity of the pronunciaiton leads to a natural lengthening of the duration of the sound, and that is why strong [consonants] differ from weak ones by greater length. [However,] the adjoining of two single weak sounds does not produce a strong one […] Thus, the gemination of a sound does not by itself create its tension." Nonetheless, Ladefoged and Maddieson examined Kodzasov's Archi recordings, and their impression was that "length should be given the primary role; strong consonants have approximately twice the duration of weak ones, and they often do result from adjoining two single consonants, at least morphologically speaking. The patterns in other Dagestanian languages are similar, but some Agul dialects have an especially large number of permitted initial long consonants." Fortis stops in Australian languages such as Rembarrnga also involve length, with short consonants having weak contact and intermittent voicing, and long consonants full closer, a more powerful release burst, and no voicing. It is not clear if strength makes the consonants long, or if during long consonants there is greater opportunity for full articulation.
Articulatory strength can reinforce other distinctions. The Ewe language, for example, which contrasts a voiceless bilabial fricative /ɸ/ and a voiceless labiodental fricative /f/, pronounces the /f/ markedly more strongly than /f/ in most languages, in contrast with the weaker /ɸ/. This helps differentiate what would otherwise be an exceedingly subtle distinction. Phonetically, a diacritic from the Extensions to the IPA can be used to indicate this strong articulation: [ɸ] vs. [f͈].
In the Mixe-Zoquean language Mixe the distinction between consonants described as fortis and lenis has been demonstrated to be one of quantity: fortis consonants are pronounced longer than their lenis counterparts, and they are also not prone to voicing in voiced environments such as the lenis consonants are.