Spoken language is a dynamic system that consists, inter alia, of connections and flows in sound production.
The primary generator of connections and flows in sound production is the prosody of language: intonation, rhythm, stress, intensity and tempo. The integrator/carrier/mirror of all these elements is intonation.
Intonation is generated by the vocal cords and is represented physically by the fundamental frequency of the voice or F0.
Intonation is responsible for the melody of the language and is inextricably connected with the flow and, therefore, temporal characteristics of spoken language. There can be no flow of language without intonation and its temporal organization.
Intonation is always present whenever any spoken language is produced (even if people commonly say that language has no intonation, it actually, and automatically, does have an F0).
Intonation is ever-present in spoken language and, therefore, can be considered to be a global phenomenon. It is literally everywhere.
F0 is necessarily a part of every vowel and voiced consonant, i.e. the vast majority of the sounds of language. This can be verified by acoustic analysis.
As a consequence, no vowel or voiced consonant can be produced without it being integrated into an intonative context.
Thus, individual phonemes never exist in isolation. They are all embedded in an intonative context – even single phonemes (the vast majority) contain a component of F0 and the others are enveloped by the intonation of the speech stream.
When intonation changes, it affects the quality (perception and production) of the vowels and consonant sounds inevitably connected with it.
As a result of this close connection, the intonation (embodying all prosodic features of spoken language) will share its tensions, pitches, flow and other pronunciation characteristics with the phonemes of the language being spoken. Simultaneously, the vowels and consonants have an impact on the production of intonation but with much less impact as they are not omnipresent.
Given its continuous nature and omnipresence, intonation will produce countless melodic variations and contexts for each individual sound to be realized in and the sounds produced in these varied contexts will automatically adjust to these melodic variations thus ultimately resulting in learners (both L1 and L2) perceiving and appropriating the acceptable variations of phoneme production for each sound of language.
Developing the above, because intonation is a holistic, ever-present, non-phoneme-specific, phenomenon of spoken language it necessarily embodies all of the pronunciation and phonological characteristics of the language as socially distributed and commonly practiced in the speech community of the language being spoken (including such phenomena as tension, loudness, stress etc.).
Thus, pronouncing individual sounds in specific intonative contexts reflects the optimal pronunciation of these sounds as they are distributed along the intonative curve according to the principles applying to the language being spoken. This helps them to be pronounced in a socially acceptable way according to the rules of the speech community of the language being spoken.
Therefore, in the context of L2 learning, listening to and pronouncing phonemes within correct intonative contours will help achieve correct/acceptable/intelligible pronunciation as the qualities of the intonation will imprint themselves on the sounds being produced and assist in their perception and production. To expand slightly on this, intonation imprints the general perceptual and pronunciation characteristics of the language on the vowels and consonants to which it is connected. As a result, it eliminates or reduces misarticulation errors of the vowels and consonants of the language being learned as it envelops these vowels and consonants with the pronunciation characteristics that they are meant to contain and automatically transfers these to them.
Further, given the generalizability of the intonative phenomenon, applying correct intonation to sounds that have never yet been uttered will also help with their production as the intonative phenomenon is applied to all speech production.
Thus, from a language learning perspective, each sound is learned as a dynamic model, or as a range of possible realizations governed by intonative contexts and variations, rather than as a static, artificially isolated, model. This is in major contradiction with common practice in pronunciation training around the world.