It is generally admitted that the sound sequence ^*μγα(*üge) has developed in Khalkha Mongolian into the two long vowels: o (ö) and a (e). Different explanations have been given as to the cause of this divergent development.
Vladimirtsov regarded the contraction *μγα>a (*üge>e) as a general development and *μγα>o (*üge>ö) as the result of progressive labial attraction. This, however, cannot be applied to such Khalkha words as dζolo «rein», dζoro «amble» etc. which originally had the vowel *i of the first syllable. Because, if we take into consideration the Ordus forms dli15 «rein» and dzilo «amble», it is most brobable that the first vowel of Khalkha dzolo and dzcoro has been brought about by the breaking of *i into o under the influence of the following o.
On the other hand, Poppe and Thomsen considered the con-traction*μγα>o (*üge>ö) to be a general development, which is, in its turn, inconsistent with other vowel constractions in Mongolian: e. g.*αγα>a, *αγu>u, *uγu>u, *oγα>o etc. As a result, they had either to take *uγa> a (*üge >e) for an ex-ception to the sound change or to assume a roundabout develop-ment such as *αburad>*abod>awid «having taken».
The present author assumes at the early stage of development the sound change * uγa>*oγa (*üge>*öge) after the vowels *o (*ö) and *i of the first syllable, only by way of which we can accept the followings as regular sound changes in Mongolian:
1. the vowel contractions *uγa>a, *üge>e, *oγa>o, *öge>ö,
2. the breaking of *i into o (ö) before o (ö).
The long vowel a (e) in Kalmuck (e. g. dzola «rein», dlora «amble») should be explained either by the relaxation of the rounded vowel o (ö) or by the contraction *oγa>a (*öge>e) peculiar to this dialect.

View full abstract

This paper attemps to make clear under what conditions the antecedent of a sentential anaphor includes the negative of the preceding (succeeding) sentence and under what conditions it does not.
Section I shows that this kind of contrast results from the difference in intersentential semantic connection, and proposes the idea of forward/backward connection. Section II elaborates the case of backward connection and proposes the following hypothesis:
(1) In the case of forward connection, the negative is included. In the case of backward connection, if a [-contrast] connective and a [+negative] predicate are used, or if a [+contrast] connective and a [-negative] predicate are used, the negative is excluded. Otherwise, it is included.
There are two exceptions to hypothesis (1). Section III takes up the one case, where a ‘negative implicative’ called by Karttunen (1971) is used in a sentence with an anaphor. It is shown that in such a case the negative never fails to be excluded from the antecedent and why this is the case.Section IV deals with the other exceptional case, where alternative connectives such as or, otherwise are used. In this case, too, the negative is always excluded from the antecedent, and the reason is made clear.

View full abstract

Although the claim has been repeated that some version of “sound hierarchy” is indispensable in the field of phonology, no one has so far offered enough data to prove it.
I conducted the statistical investigations to support explicitly the order of “Japanese sound hierarchy” using the artificial onomatopoetic words of AB-type. 582 college students who were mainly from Western Japan were asked if they could find a preferred order in“ pira-pura vs. pura-pira, ” for example. Chi-square test was used to certify the significance levels of all the minimally contrastive pairs that were grouped into 87 terms. Then, the results were combined to obtain the following.
Vowel Hierarchy: i<u<e<a<o
Consonant Hierarchy: glides<liquids<nanals<fricatives<stops<affricates
In addition, the double-function system was introduced to some exceptional sounds in the consonant hierarchy.

View full abstract