Fusion

id	Language	Cumulation of TAME and case	Morphologically conditioned stem alternation: suppletion	Morphologically conditioned stem alternation: Iregular
1	Bantawa	Bantawa has an extensive predicate marker paradigm, in which portmanteaus combine person, number, tense and polarity, e.g. man-…daca ‘NEG.PST.1DU’. (Doornenbal 2009: 145, 147ff.). Case is marked by affixes that do not express other semantic categories.	There is no morphologically conditioned stem alternation in Bantawa, such that it would result in a many-to-one relation between meanings and stems.	There is no morphologically conditioned stem alternation in Bantawa, such that it would result in a many-to-one relation between meanings and stems.
2	Bininj Gun-Wok	There is cumulation of tense, person, and number in pronominal verbal prefixes, e.g. bi- ‘3>3.PST’ (Evans 2003: 399ff.). Case markers, called ‘role affixes’ by Evans (cf. Evans 2003: 136ff.) do not cumulate with other categories.	The morphology of Bininj Gun-Wok is strongly agglutinative, showing only few morphological or morphophonemic alternations, leading Evans (2003: 106) to call the language ‘lego-like’. However, there is one suppletive verb, viz. re ‘go’, that inflects for TAME by means of stem change: ray ‘go.IMP’, re ‘go.NPST’, wam ‘go.PST.PFV’, etc. (Evans 2003: 361).	Bininj Gun-Wok shows a case of irregular stem formation, viz. the verb wokdi ‘speak’ is irregular, as it inflects a little differently from other verbs in its verbal class (viz. conjugation 7, cf. Section 3.4.7): the past perfective is in that class marked by means of a zero-suffix, but wokdi undergoes a stem change and becomes wokdanj ‘speak.PST.PFV’ (Evans 2003: 350).
3	Chukchi	There is cumulation of tense and polarity in negation particles, e.g. wanewan ‘NEG.NFUT’, qərəm~qəcəm ‘NEG.FUT’ (Dunn 1999: 326, 327). Absolutive case is cumulated with singularity (Dunn 1999: 111).	There is no suppletion of stems in Chukchi.	Morphologically conditioned stem alternation is observed in the marking of absolutive singular (Dunn 1999: 105ff.). There are five strategies to do this, two of them involving stem alternation. The first stem alternation strategy consists of a reduction or deletion of the stem-final vowel, e.g. wala-t ‘knife.ABS-PL’, walə ‘knife.ABS.SG’ (Dunn 1999: 106). This stem alternation is phonologically conditioned, as it only occurs with stems ending in e or a and is completely predictable. However, the second stem alternation strategy involves irregular and unpredictable changes to stems, e.g. welw-ət ‘raven.ABS-əPL’, wetlə ‘raven.ABS.SG’ (Dunn 1999: 109). While plurality is marked by means of a suffix, singularity cannot be said to be marked by a zero-morpheme, since the stem undergoes an alternation if singular. Therefore, the stem expresses both its lexical meaning and singularity. Since this process applies only in absolutive case, the stem expresses case information as well.
4	Dutch	In some Dutch pronouns, case marking is cumulated with person and number marking, e.g. hij ‘NOM.3SG.M’, hem ‘ACC.3SG.M’. Marking of TAME on verbs is cumulated with marking of person and number, e.g. -te ‘-PST.SG’.	Suppletion occurs in Dutch in the paradigms of certain verbs, e.g. zijn ‘be.INF’, ben ‘be.PRS.1SG’, is ‘be.PRS.3SG’, zijn ‘be.PRS.3PL’, was ‘be.PST.SG’.	Irregular stem formation in the form of ablaut appears in Dutch in the tense forms of strong verbs, e.g. geef ‘give.PRS.1SG’, gaf ‘give.PST.SG’.
5	Egyptian Arabic	In Egyptian, basic tense and aspect distinctions are expressed by means of the CV-patterns that are inserted between consonantal roots, e.g. ji-ktib ‘3SG-write.PRS.IPFV’, he is writing’, katab-Ø ‘write.PST.PFV-3SG, he wrote’. These tense and aspect markers cannot be seen separately from person and number marking because, as demonstrated in these examples, there are different sets of person and number marking affixes dependent on the TAME specification (cf. Gary & Gamal-Eldin 1981: 100). Therefore, we can say that Egyptian shows cumulation of TAME, person and number. There are also tense and aspect markers that do not show cumulation with person and number; these are affixes and auxiliaries (Gary & Gamal-Eldin 1981: 92). There is no case marking in Egyptian.	Egyptian exhibits suppletion in the formation of the imperative of some verbs, e.g. ana geet ‘1SG come.IND.1SG’, ta?ali ‘come.IMP’, in which the consonantal roots are altered in imperative mood (L. C. Buell, personal communication, November 1, 2013).	As explained in Section 6.2.4, Egyptian Arabic makes use of so-called transfixes: inflectional affixes consisting of a particular CV-pattern, that are integrated with three-consonantal roots to form a stem. This type of morphology is sometimes analysed as stem alternation, since the stem vowels change, resulting in a form that represents multiple meanings, for example the lexical meaning of a verb and tense and aspect operators in ji-ktib ‘3SG-write.PRS.IPFV, in one inseparable form. However, in agreement with many Arabists, I see the three consonants as the true stem, while the vowels in between are the inflectional affixes, i.e. transfixes. Since the stem, i.e. the consonantal roots, do not undergo modification under TAM and person/number inflection, I do not see this as stem alternation. However, there are true cases of morphologically conditioned stem alternation in Egyptian Arabic, affecting the consonantal roots, e.g. irregular stem formation of verbal stems. For example, causative is expressed by means of gemination of the second root, which leads to a double consonant, e.g. for the consonantal roots FHM, we get fihim ‘to understand’, fahhim ‘understand.CAUS, to make clear’ (Woidich & Heinen-Nasr 2000: 154). Since the causative meaning is inseparable from the lexical meaning of the stem, this qualifies as fusional stem alternation.
6	Fongbe	Fongbe has several TAME markers (Lefebvre & Brousseau 2002: 89), which are not cumulated with other categories, but it also has some TAME marking on particular personal pronouns, e.g. ná ‘FUT.1SG’, màá ‘NEG.FUT.1SG’ (E. Aboh, personal communication, July 9, 2014). It also has two case markers (Lefebvre & Brousseau 2002: 44), which are not cumulated with other categories.	I have not found examples of morphologically based stem alternation in Fongbe.	I have not found examples of morphologically based stem alternation in Fongbe.
7	Georgian	Third person pronouns exhibit cumulation of case, person, and number, e.g. is ‘NOM.3SG’, man ‘ERG.3SG’ (Vamling 1989: 17). Furthermore, tense, aspect and person are cumulated in portmanteau predicate suffixes, e.g. –s ‘PRS.3’, -a ‘AOR.3’ (Vamling 1989: 18).	There is extensive stem alternation in Georgian. For instance, particular verbal stems mark plurality by means of suppletion, viz. zi-s ‘s/he is seated’, sxed-an ‘they are seated’ (Hewitt 1995: 214).	There is extensive irregular stem formation in Georgian, for example in some verbs, on which animacy of the Undergoer is marked by means of irregular stem formation, e.g. in example (8). Hewitt (1995: 214) (8) a. c’ign-i m-a-kv-s book-NOM 1SG-LV-have.INAN.U-PRS.3 ‘I have a/the book.’ b. da m-q’av-s sister.NOM 1SG-have.ANIM.U-PRS.3 ‘I have a sister.’
8	Huallaga Quechua	There is no cumulation of case and another category in Quechua, cf. Weber (1989: 55). Person and tense are marked on verbs by means of so-called transition complexes (cf. Weber 1989: 95ff.), which are predicate markers consisting of an object marker, a tense marker and a subject marker, in that order. Weber analyses them as entities but in fact, they can be separated in multiple suffixes that each express one meaning, e.g. -man could be analysed as -ma-Ø-n ‘1.OBJ-PRS-SBJ’ and -maran as -ma-ra-n ‘1.OBJ-PST-SBJ’. This becomes harder with future tense, which is zero-marked for certain combinations of subject and object person, and with first person plural inclusive suffixes, which undergo alternations, but still, there is, in my opinion, no cumulation.	I have not found instances of suppletion in Quechua.	I have not found instances of irregular stem formation in Quechua.
9	Japanese	TAME specifications in Japanese are expressed by means of verbal affixes, but not cumulated with other categories (Iwasaki 2002: 105ff.). Case and pragmatic function are expressed by means of morphosyntactically free elements, and are also not cumulated with other categories (Iwasaki 2002: 44).	I found no instances of morphologically based stem alternation in Japanese; all alternation processes are determined (morpho)phonologically.	I found no instances of morphologically based stem alternation in Japanese; all alternation processes are determined (morpho)phonologically.
10	Kayardild	Kayardild is an agglutinative, suffixing language (Evans 1995: 1). Neither case nor TAME suffixes cumulate with other categories. Of course, as discussed in Section 11.1.2, there are the modal case markers that express modality, which could be seen as markers of both case and mood. However, these are not ‘case markers’ in the normal sense of that word, since they do not mark grammatical relations (nominative, accusative) or locational (ablative, locative) information. Rather, they are mood markers that are diachronically related to regular case markers, but are not case markers themselves.	There is no suppletive stem alternation in Kayardild.	Middle voice is expressed in Kayardild by means of the suffix -yii and by alternations in the stem. On verbal stems ending in a long vowel, middle voice is accompanied by a shortening of that vowel, e.g. baa-ja ‘bite-INF’, ba-yii-ja ‘bite-MID-INF’ (Evans 1995: 277). In other verb stems, the vowel is lengthened: diya-ja ‘eat-INF’, diyaa-ja ‘eat.MID-INF’. The latter example shows that -yii only appears with some verbal stems, but not with others. This means that in some stems, e.g. diyaaja, middle voice is expressed only by means of stem alternation.