This continues the discussion of the ezāfe begun in previous two sections.

If in possessive construction, after the silent /h/ the personal modifiers are used, the spoken inflection would include an anaptictic /^ə/ (not a full grade /a/) between the final /-e/ sound (which is represented by the silent /h/ in written form) and the plural endings, as follows:

In a variation, the final /e/ is assimilated into the initial /a/ of the postvocalic enclitics, as follows:

After a final long /ī/ ‍ی (here we use i), however, we do not need an euphonic /y/ to help us with the articulation. Hence, in صندلی معلّم ما sandali-ye mo’allem-e mā ‘our teacher’s chair,’ the [–e] alone would suffice. (It should be understood that the [-y-] in sandali-ye mo’allem is only necessary in the English transcription.) There is a debate over the latter case. Some Persian scholars prefer to use the euphonic /y/ even after a terminal ی (for example, صندلی ی من sandali-ye man), mostly for the sake of consistency.

After the silent /h/, sometimes instead of a full grade ی a diacritic is placed on the ـه in the writing, for instance, خانه‌ٔ من and خانه ‌ی من both pronounced xāne-ye man ‘my house.’

It must be understood that this character is not a hamze; it is the development of the letter ی as a diacritic, due to the frequency of usage. Its design is simply taken from the upper portion of the letter ی, which, for the sake of fluency, does not curve upwards and back, but is left in the middle of the character in such a way that flowing forward is made smoother. Even referring to this diacritic as hamze is incorrect; it is an ezāfe ی (ye-ye ezāfe, or yā’-e ezāfe).

Since in Middle Persian the ezāfe ی was written as a separate character, there are some scholars who prefer to incorporate a full grade ی; for example, بچّه ‌ی من bace-ye man ‘my child’. This is a sort of deconstruction and reviving of the historical form. Both بچّه ‌ی من and بچّهٔ من are correct, and the usage of one vs. the other is only a matter of choice.

However, the incorporation of this yā’-e ezāfe for the indefinite ی by some copyists throughout history is entirely erroneous, and merely based on the north-eastern phonetic influence. That is to say, in the north-eastern dialects (such as Dari) there is not a distinct difference of articulation between the ezāfe ی and the indefinite enclitic ی; for example, xāna-yi dāram ‘I have a house’ vs. xāna-yi man ‘my house.’ That is why scribes have confused the two. As a result, in some of the classical and archaic Persian texts one may see خانهٔ من ‘my house’ and خانهٔ دارم ‘I have a house,’ both with [-yi]. The latter, of course, is incorrect; simply because in the indefinite case, after the silent /h/, we would have an euphonic ی intervening as یی (for example, خانه ‌یی دارم ) while in the ezāfe there is only one ی (for example, خانه ‌ی من ).

Names of streets, seas, oceans, seasons, weeks, mountains, universities, etc., fall into this category and take the ezāfe linking [-e]:

خیابان فردوسی
xiyābān-e ferdowsi
‘the Ferdowsi Street’

دریای سیاه
daryā-ye siyāh
‘The Black Sea’
(notice the euphonic [ye] after the long /ā/)

اقیانوس آرام
oqyānus-e ārām
‘The Pacific Ocean’

فصلِ بهار
fasl-e bahār
‘the spring season’

هفتهٔ دیگر
hafte-ye digar
‘next week’
(notice the euphonic [ye] after the silent /h/)

کوه البرز
kuh-e alborz
‘The Alborz Mountain’

کوه آلپ
kuh-e ālp
‘The Alps’

دانشگاه تهران
dānešgāh-e tehrān
‘The Tehran University’

In addition, the ezāfe corresponds to the English of in such constructions as ‘The City of Berkeley,’ ‘the third week of the month,’ etc.; for example, شهر تهران šahr-e tehrān ‘The City of Tehran’; and it is even used with the name of countries; for example, کشور ایران kešvar-e irān ‘Iran’, کشور آمریکا kešvar-e āmrikā ‘America, the US of A’; literally ‘the land of Iran,’ ‘the land of America’ (although کشور kešvar = ‘country’), هفتهٔ دوم سال hafte-ye dovom-e sāl ‘the second week of the year.’ etc.

When dealing with a chain of attributive adjectives, the ezāfe occurs between all of the adjectives, while all the euphonic features are also observed:

خانهٔ بزرگِ قدیمی سفید
xāne-ye bozorg-e qadimi-ye sefid
‘the big, old, white house’

If the conjunction /و/ ‘and’ (and its replacement, the comma /،/) is used between
adjectives, then the ezāfe is suffixed only at the end of the last adjective:

خانهٔ بزرگ، قدیمی و سفیدِ ما
xāne-ye bozorg, qadimi va sefid-e mā
‘our big, old, white house’

(Notice that in English using the conjunction and before “white” is not necessary.)

In a combination of attributive adjectives and possessives, still the ezāfe occurs between all the adjective and the possessive combinations:

خانهٔ بزرگِ قدیمی سفیدِ ما
xāne-ye bozorg-e qadimi-ye sefid-e mā
‘our big, old, white house’

خانهٔ، بزرگ، قدیمی و سفیدِ پدر ما
xāne-ye bozorg, qadimi va sefid-e pedar-e mā
‘our father’s big, old, white house’

In the Persian transcription of the ezāfe construction never use a final [ه/ﻪ] instead of the ezāfe enclitic [-e] (the kasre). That is to say, never write خواهره من or کتابه خوب ; but always خواهرِمن or کتابِ خوب .

It must be understood that hamze /ء/ is a consonant. Therefore, in the ezāfe construction, a final hamze takes the regular [-e] (and not [-ye]). In other words,[ezafe3_ 50] انشاءِ من enšā’-e man ‘my composition’ is correct and not انشای من.In the Persian mutation of the Arabic loanwords, however, the final hamze of many such words has been dropped and a second variation of such words exists (only in Persian); namely, اعضاء a’zā’ ~ اعضا a’zā ‘members’, آراء ārā’ ~ آرا ārā ‘votes’, etc.

In this case, since we are dealing with a final /ā/, naturally we need to use the euphonic [-y-] in the ezāfe construction, as اعضاءِ بدن a’zā’-e badan vs. اعضای بدن a’zā-ye badan ‘body members (i.e., limbs)’, آراءِ ما ārā’-e mā vs. آرای ما ārā-ye mā ‘our votes’, etc.