Table 2. Ordering relations among grammatical morphemes
______________________________________________________________________

1.    Clear cases (true at least 90 per cent of the time. Data from 
    studies where each morpheme occurs in at least ten obligatory 
    occasions. Counterexamples discussed in Table 3.)
     1. ING precedes Irr. PAST
     2. ING precedes Reg. PAST
     3. ING precedes III singular
     4. ING precedes POSS
     5. ING precedes AUX
     6. PLURAL precedes Ir. PAST
     7. PLURAL precedes Reg. PAST
     8. PLURAL precedes III singular
     9. PLURAL precedes POSS
    10. COP precedes Irr. PAST
    11. COP precedes Reg. PAST
    12. COP precedes III singular
    13. COP precedes POSS
    14. ART precedes POSS
    15. ART precedes III singular
    16. ART precedes Reg. PAST
    17. AUX precedes POSS
    18. AUX precedes III singular
    (1 through 9 hold for all studies; 10 through 18 do not necessarily
    hold for child L1 acquisition)

II.   Less clear cases (true of most studies where each morpheme occurs
    in at least ten obligatory occasions. Except where indicated, 
    probability of relation holding significant by sign test.)
     1. ING precedes ART
     2. ING precedes PLURAL
     3. PLURAL precedes AUX
     4. PLURAL precedes ART (p < 15)
     5. COP precedes AUX
     6. COP precedes ART
     7. ART precedes Irr. PAST
     8. AUX precedes Irr. PAST
     9. AUX precedes Reg. PAST
    10. Irr. PAST precedes III singular
    11. Irr. PAST precedes Reg. PAST (p < 0.21)
    12. Irr. PAST precedes POSS (p < 0.11)
    (Numbers 1, 3, and 7 hold for child L1 acquisition; the others do not
    necessarily hold for child L1 acquisition.)
______________________________________________________________________

are not given, or come from Uguisu, Hakuta's subject. (Detailed discussion of Uguisu will wait for another study.)

Many of the generalizations hold for L1 as well as L2. In general, this is true of bound morphemes, and this agreement underlies the

58