Building blocks of rhythmic speech
Every syllable contains a nucleus, typically a vowel or vowel-like sound that forms its core. Most syllables also have an onset, one or more consonants preceding the nucleus, and a coda, one or more consonants following the nucleus. The word 'strong' has the structure CCVCCC: two consonants in the onset, one vowel as the nucleus, and three consonants in the coda. The word 'at' is a VC syllable with no onset. The word 'I' is just a nucleus with no onset or coda. Most languages follow similar principles, though they vary dramatically in how many consonants they allow in onsets and codas.
The nucleus can be filled by sounds other than vowels. In English, syllabic consonants like the 'l' in 'bottle' can form the nucleus of a syllable, though they are less common than vocalic nuclei.
Cross-linguistic constraints and phonotactics
Languages impose constraints on which consonants can appear together. English allows 'str' at the beginning of a syllable but not 'tsr'. Japanese generally does not allow consonant clusters in onsets or codas, making every syllable approximately CV or V. These phonotactic constraints are part of a language's grammar and are learned by children implicitly. The constraints reflect both universal principles about articulatory ease and arbitrary language-specific conventions passed through generations.