Canonical and Non-Canonical Roots: The Diversity of Roots in Mandarin Chinese

Qin defines a root as a morphologically unanalyzable form that serves as a core of word formation, the substrate's Morphology.IsCore over Bloomfield's primary words. The judgment is formal rather than semantic: the meaningless -fer is a core of the primary word refer, and the plural -s of cats is not. Four criteria, boundness, positional flexibility, phonological richness and meaning lexicality, span sixteen cells, $\binom{4}{k}$ of which fail $k$ criteria. Each of 122 Mandarin morphemes falls in the group of the number of criteria it fails, the groups run from the canonical 厂 chǎng to the toneless 个 -ge, and a toneless morpheme fails all four.

Implementation notes

The appendix's codes are the features of each form's cell in the data: F or B for boundness, FLEX, SL, or SR for position, 1 or TL for phonology, and L or NL for meaning. A row reads them as a Boolean judgment on each criterion, the canonical value being true; the two fixed positions SL and SR both count as fixed. The appendix's group is the paper's analysis of the form. The counts of Tables 2 and 4 are the distribution of the forms and are not restated.

TODO

The paper says that six of the unattested cells of the space require phonological deficiency; the appendix leaves seven such cells unattested, since only the maximally non-canonical cell contains a toneless morpheme, and the cell of free, fixed, toned, lexical morphemes is unattested as well.

References

• Qin 2025, Canonical and Non-Canonical Roots: The Diversity of Roots in Mandarin Chinese
• Bloomfield 1933, Language formalized