The main difference between a tap or flap and a stop is that in a tap/flap there is no buildup of air pressure behind the place of articulation and consequently no release burst. Otherwise a tap/flap is similar to a brief stop.
Taps and flaps also contrast with trills, where the airstream causes the articulator to vibrate. Trills may be realized as a single contact, like a tap or flap, but are variable, whereas a tap/flap is limited to a single contact. When a trill is brief and made with a single contact it is sometimes erroneously described as an (allophonic) tap/flap, but a true tap or flap is an active articulation whereas a trill is a passive articulation. That is, for a tap or flap the tongue makes an active gesture to contact the target place of articulation, whereas with a trill the contact is due to the vibration caused by the airstream rather than any active movement.
Many linguists use the terms tap and flap indiscriminately. Peter Ladefoged proposed for a while that it might be useful to distinguish between them. However, his usage was inconsistent and contradicted itself even between different editions of the same text. One proposed version of the distinction was that a tap strikes its point of contact directly, as a very brief stop, but a flap strikes the point of contact tangentially: "Flaps are most typically made by retracting the tongue tip behind the alveolar ridge and moving it forward so that it strikes the ridge in passing."
Later, however, he used the term flap in all cases. Subsequent work on the labiodental flap has clarified the issue: flaps involve retraction of the active articulator, and a forward-striking movement.
For linguists who make the distinction, the alveolar flap is transcribed as a fish-hook ar, [?], and the tap can be transcribed as a small capital D, [?], which is not recognized by the IPA, or by [d?]. In IPA terms the retroflex flap [?] symbol captures the initial retraction and subsequent forward movement of the tongue tip involved. Otherwise, alveolars are typically called taps, and other articulations are called flaps.
A few languages have been reported to contrast a tap and a flap at the same place of articulation. This is the case for Norwegian, in which the alveolar apical tap /?/ and the post-alveolar/retroflex apical flap /?/ have the same place of articulation for some speakers, and Kamviri, which also has apical alveolar taps and flaps.
The tap and flap consonants identified by the International Phonetic Alphabet are:
|?||alveolar tap||North American English||latter||[læ??]||"latter"|
|?||alveolar lateral flap||Venda||vula||[vu?a]||"to open"|
The Kiel Convention of the IPA recommended that for other taps and flaps, a homorganic consonant, such as a stop or trill, should be used with a breve diacritic:
Tap or flaps: where no independent symbol for a tap is provided, the breve diacritic should be used, e.g.  or [n?].
However, the former could be mistaken for a short trill, and is more clearly transcribed ⟨⟩, whereas for a nasal tap the unambiguous transcription ⟨⟩ is generally used.
|Central oral||(b?, w?)||(f?) · ? (v?)||· ?||· ?||()||()|
|Central nasal||(m?)||(n?)||( )|
|Lateral oral|| · ?||? ( )|
Most of the alternative transcriptions in parentheses imply a tap rather than flap articulation, so for example the flap  and the tapped stop [b?] are arguably distinct, as are flapped  and tapped .
Spanish features a good illustration of an alveolar tap, contrasting it with a trill: pero /'pe?o/ "but" vs. perro /'pero/ "dog". Among the Germanic languages, this allophone occurs in American and Australian English and in Northern Low Saxon. In American and Australian English it tends to be an allophone of intervocalic /t/ and /d/, leading to homophonous pairs such as "metal" / "medal" and "latter" / "ladder" - see Flapping. In a number of Low Saxon dialects it occurs as an allophone of intervocalic /d/ or /t/; e.g. bäden /beeden/ -> ['be:?n] 'to pray', 'to request', gah to Bedde! /gaa tou bede/ -> [:to?'be?e] 'go to bed!', Water /vaater/ -> ['v?:] 'water', Vadder /fater/ -> ['fa] 'father'. (In some dialects this has resulted in reanalysis and a shift to /r/; thus bären ['be:rn], to Berre [to?'bere], Warer ['v?:r?], Varrer ['far?].) Occurrence varies; in some Low Saxon dialects it affects both /t/ and /d/, while in others it affects only /d/. Other languages with this are Portuguese, Korean, and Austronesian languages with /r/.
Most Indic and Dravidian languages have retroflex flaps. In Hindi there are three, a simple retroflex flap as in [b:] big, a murmured retroflex flap as in [koi:] leper, and a retroflex nasal flap in the Hindicized pronunciation of Sanskrit [mi] ruby. Some of these may be allophonic.
Many of the languages of Africa, Asia, and the Pacific that do not distinguish [r] from l may have a lateral flap. However, it is also possible that many of these languages do not have a lateral-central contrast at all, so that even a consistently neutral articulation may be perceived as sometimes lateral [?] or [l], sometimes central [?]. This has been suggested to be the case for Japanese, for example.
The Iwaidja language of Australia has both alveolar and retroflex lateral flaps. These contrast with lateral approximants at the same positions, as well as a retroflex tap [?], alveolar tap [?], and retroflex approximant [?]. However, the flapped, or tapped, laterals in Iwaidja are distinct from 'lateral flaps' as represented by the corresponding IPA symbols (see below). These phones consist of a flap component followed by a lateral component, whereas In Iwaidja the opposite is the case. For this reason, current IPA transcriptions of these sounds by linguists working on the language consist of an alveolar lateral followed by a superscript alveolar tap and a retroflex lateral followed by a superscript retroflex tap.
for this sound. (Supported by some fonts: [?].) Previously, it had been transcribed with the use of the breve diacritic, [v?], or other ad hoc symbols.
Other taps or flaps are much less common. They include an epiglottal tap; a bilabial flap in Banda, which may be an allophone of the labiodental flap; and a velar lateral tap as an allophone in Kanite and Melpa. These are often transcribed with the breve diacritic, as [w?, ]. Note here that, like a velar trill, a central velar flap or tap is not possible because the tongue and soft palate cannot move together easily enough to produce a sound.
If other flaps are found, the breve diacritic could be used to represent them, but would more properly be combined with the symbol for the corresponding voiced stop. A palatal or uvular tap or flap, which unlike a velar tap is believed to be articulatorily possible, could be represented this way (by *[, ~]).
Nasalized consonants include taps and flaps, although these are rarely phonemic. Many West African languages have a nasal flap  (or [n?]) as an allophone of /?/ before a nasal vowel; Pashto, however, has a phonemic nasal retroflex lateral flap.
Voiced and voiceless tapped alveolar fricatives have been reported from a few languages. Flapped fricatives are possible but do not seem to be used. See voiced alveolar tapped fricative, voiceless alveolar tapped fricative.