[2] The minimal pair was an essential tool in the discovery process and was found by substitution or commutation tests.
Sometimes, the phenomenon can create some syntactic-gemination-minimal-pairs: In the example, the graphical accent on dà is just a diacritical mark that does not change the pronunciation of the word itself.
For example, Kono, of Sierra Leone, distinguishes high tone and low tone on syllables:[4][5] Languages in which stress may occur in different positions within the word often have contrasts that can be shown in minimal pairs, as in Greek and Spanish: In English stress can determine the part of speech of a word: insult as a noun is /ˈɪnsʌlt/ while as a verb it is /ɪnˈsʌlt/.
At the word boundary, a "plus juncture" /+/ has been posited and said to be the factor conditioning allophones to allow distinctivity:[7] in this example, the phrase "great ape" has an /eɪ/ diphthong shortened by pre-fortis clipping and, since it is not syllable-initial, a /t/ with little aspiration (variously [t˭], [ɾ], [ʔt], [ʔ], etc., depending on dialect); meanwhile in "grey tape", the /eɪ/ has its full length and the /t/ is aspirated [tʰ].
[10] For example, the vowels /a/, /e/, /i/, /o/, /u/ of Swahili are shown to be distinct by the following set of words: pata 'hinge', peta 'bend', pita 'pass', pota 'twist', puta 'thrash'.
[11] However, establishing such sets is not always straightforward [12] and may require very complex study of multiple oppositions as expounded by, for example, Nikolai Trubetzkoy.
For example, if the focus of a lesson was on the distinction /ɪ/ versus /ɛ/, learners might be asked to signal which sound they heard as the teacher pronounced lists of words with these phonemes such as lid/led, tin/ten, or slipped/slept.
[15] However, even today minimal pair listening and production drills remain a common tool for the teaching of segmental differences.
[20] Signs consist of phonemes, which are specifications for location, movement, handshape, orientation, and non-manual elements.