The pronunciation of Modern Standard Chinese is based on the Beijing dialect, which is of the Northern, or Mandarin, type. It employs about 1,300 different syllables. There are 22 initial consonants, including stops (made with momentary, complete closure in the vocal tract), affricates (beginning as stops but ending with incomplete closure), aspirated consonants, nasals, fricatives, liquid sounds (l, r), and a glottal stop. The medial semivowels are y (i), ɥ (ü), and w (u). In final position, the following occur: nasal consonants, ṛ (retroflex r), the semivowels y and w, and the combinations ŋr (nasalization plus r) and wr (rounding plus r). There are nine vowel sounds, including three varieties of i (retroflex, apical, and palatal). Several vowels combine into clusters.

There are four tones: (1) high level, (2) high rising crescendo, (3) low falling diminuendo with glottal friction (with an extra rise from low to high when final), and (4) falling diminuendo. Unstressed syllables have a neutral tone, which depends on its surroundings for pitch. Tones in sequences of syllables that belong together lexically and syntactically (“sandhi groups”) may undergo changes known as tonal sandhi, the most important of which causes a third tone before another third tone to be pronounced as a second tone. The tones influence some vowels (notably e and o), which are pronounced more open in third and fourth tones than in first and second tones.

A surprisingly low number of the possible combinations of all the consonantal, vocalic, and tonal sounds are utilized. The vowels i and ü and the semivowels y and ɥ never occur after velar sounds (e.g., k) and occur only after the palatalized affricate and sibilant sounds (e.g., tś), which in turn occur with no other vowels and semivowels.

Many alternative interpretations of the distinctive sounds of Chinese have been proposed; the interaction of consonants, vowels, semivowels, and tones sets Modern Standard Chinese apart from many other Sinitic languages and dialects and gives it a unique character among the major languages of the world. The two most widely used transcription systems (romanizations) are Wade-Giles (first propounded by Sir Thomas Francis Wade in 1859 and later modified by Herbert A. Giles) and the official Chinese transcription system today, known as the pinyin zimu (“phonetic spelling”) or simply Pinyin (adopted in 1958). For a comparison of these romanization equivalents, see the table. In Wade-Giles, aspiration is marked by ’ (p’, t’, and so on). The semivowels are y, yü, and w in initial position; i, ü, and u in medial; and i and u (but o after a) in final position. Final retroflex r is written rh. The tones are indicated by raised figures after the syllables (¹, ², ³, ⁴).

The Pinyin system indicates unaspirated stops and affricates by means of traditionally voiced consonants (e.g., b, d) and aspirated consonants by voiceless sounds (e.g., p, t). The semivowels are y, yu, and w initially; i, ü, and u medially; and i and u (o after a) finally. Final retroflex r is written r. The tones are indicated by accent markers, 1 = ¯, 2 = ´, 3 = ˇ, 4 = ˋ (e.g., mā, má, mǎ, mà = Wade-Giles ma¹, ma², ma³, ma⁴).

Pinyin is used in the following discussion of Modern Standard Chinese grammar.

The most common suffixes that indicate nouns are -zi (as in fangzi ‘house’), and -tou (as in mutou ‘wood’). A set of postposed noun particles express space and time relationships (-li ‘inside,’ -hou ‘after’). An example of a verbal affix is -jian in kanjian ‘see’ and tingjian ‘hear.’ Important verb particles are -le (completed action), -guo (past action), and -zhe (action in progress). The directional verbal particles -lai ‘toward speaker’ and -qu ‘away from speaker’ and some verbal suffixes can be combined with the potential particles de ‘can’ and bu ‘cannot’—e.g., na chulai ‘take out,’ na bu chulai ‘cannot take out’; tingjian ‘hear,’ ting de jian ‘can hear.’ The particle de indicates subordination and also gives nominal value to forms for other parts of speech (e.g., wo ‘I,’ wode ‘mine,’ wo de shu ‘my book,’ lai ‘to come,’ lai de ren ‘a person who comes’). The most important sentence particle is le, indicating ‘new situation’ (e.g., xiayu le ‘now it is raining,’ bu lai le ‘now there is no longer any chance that he will be coming’). Ge is the most common noun classifier (i ‘one,’ yi ge ren ‘one person’); others are suo (yi suo fangzi ‘one house’) and ben (liang ben shu ‘two books’).

Adjectives can be defined as qualitative verbs (hao ‘to be good’) or stative verbs (bing ‘to be sick’). There are equational sentences with the word order subject–predicate—e.g., wo shi Beijing ren ‘I am a Beijing-person (i.e., a native of Beijing)’—and narrative sentences with the word order subject (or topic)–verb–object (or complement)—e.g., wo chifan ‘I eat rice,’ wo zhu zai Beijing ‘I live in Beijing.’ The preposed object takes the particle ba (wo da ta ‘I beat him,’ wo ba ta dale yidun ‘I gave him a beating’), and the agent of a passive construction takes bei (wo bei ta dale yidun ‘I was given a beating by him’).