Historical Izâfa Orthography and Modern Editorial Practice

Persian ezâfe has been far more stable as a syntactic linker than as a written sign. After consonants it is still normally absent from the page; after final vocalic ه, manuscripts, metal type, schoolbooks, and digital standards have represented the y of -ye in different ways. A C2 reader must distinguish a change in graphic convention from a change in grammar or pronunciation.

This page uses izâfa when discussing the historical tradition and ezâfe for contemporary Iranian Persian. The terminology is editorial convenience, not two different constructions. The modern house form is خانه‌ی من; an exact quotation may preserve خانهٔ من because the source's spelling is evidence.

For the mechanics of reading historical hands before editorial normalization, consult Reading Nastaʿlīq and Older Print Styles.

The consonant-final linker: audible, optionally marked

After a consonant, modern ezâfe is -e. Ordinary Iranian printing omits it: کتاب خوب is read ketâb-e khub. A kasre can be written under the final consonant—کتابِ خوب—in primers, dictionaries, verse annotation, ambiguous passages, and fully vocalized editions.

The two linkers are absent from target and explicit in spokenForm. Adding kasre to every consonant-final host in ordinary modern prose is not more grammatical; it changes the editorial level of vocalization.

Final vocalic ه and competing graphic traditions

After silent final ه, the linker is -ye. The historical graphic above ه is widely called “hamza” in typography, but its function here descends from a small or superscript yeh tradition; it does not represent an Arabic glottal stop. Modern sources therefore show several superficially different forms.

The first three rows need not differ in spoken Iranian Persian. The disagreement concerns encoding, glyph history, search behavior, and editorial policy. For new guide prose, this repository deliberately uses ه‌ی, with U+200C between ه and Persian ی. That house choice should not be confused with the Academy's printed دستور خط, which prescribes the superscript-sign form exemplified by خانهٔ من.

Manuscripts do not behave like normalized digital text

Persian manuscripts commonly omit short-vowel signs, vary the placement or shape of small yeh, and rely on a trained reader's lexical and metrical knowledge. A superscript mark may be faint, displaced, or visually similar to other dots. Absence of a visible kasre is therefore not evidence that the scribe intended no izâfa.

A diplomatic transcription records what can actually be seen, including uncertainty. A normalized transcription records the editor's linguistic interpretation in a consistent modern system. Neither is inherently superior; they answer different questions.

Earlier pronunciation and modern recitation

The ezâfe vowel is /e/ in standard Iranian Persian, but descriptions of earlier New Persian reconstruct or transcribe a higher vowel /i/; eastern varieties also preserve related i-like outcomes. This historical difference explains the widespread scholarly spelling izâfa. It does not license an unmarked -i pronunciation in a modern Iranian example.

The history is not read directly from one missing kasre. Manuscript spelling, comparative dialect evidence, grammatical description, and poetic phonology must be interpreted together. Meter can show that a syllable is present, but meter alone does not always determine its exact vowel quality.

In this guide's audio, a classical quotation receives a standard modern Iranian reading unless the page explicitly announces a historical reconstruction. That policy keeps learner audio coherent while allowing philological prose to discuss older -i.

Source-faithful quotation versus house normalization

Inside quotation marks, preserve the orthography of the cited edition when spelling itself matters. Outside the quotation, use modern house style and identify normalization. If an old print has خانهٔ دوست, do not silently turn it into خانه‌ی دوست in a diplomatic claim; reproduce the source form or state that you have normalized it.

This separation prevents two opposite errors: modernizing away evidence and presenting an obsolete encoding as the current recommendation.

A public-domain text and its editorial layers

Saadi's thirteenth-century Golestan contains the public-domain line بنی‌آدم اعضای یکدیگرند. A modern normalized edition may use ZWNJ in بنی‌آدم and write visible ی after final ا in اعضای. A manuscript witness may space, join, or mark those forms differently.

The quotation here is explicitly a normalized modern reading, not a diplomatic transcription of a named codex. A scholarly edition should identify its witness and transcription principles before drawing conclusions from a superscript sign.

An editor's decision sequence

First decide whether the text is a new composition, a normalized reading edition, or a documentary transcription. Next identify consonant-final and vowel-final hosts from grammar, not merely from visible marks. Then apply one convention consistently, recording every departure that is evidentially significant.

Common Mistakes

1. Reading an absent kasre as an absent linker

2. Pronouncing the superscript sign as a glottal stop

3. Using a full space in new standard prose

4. Silently modernizing a documentary quotation

5. Importing historical -i into unmarked modern speech

Key Takeaways

• Consonant-final ezâfe may be marked with kasre, but ordinary omission does not remove the spoken linker.
• The superscript sign over final vocalic ه historically represents the y boundary in this function, not a glottal stop.
• Modern guide prose uses ه‌ی; source-faithful quotation may preserve older graphic forms under a declared policy.
• Historical -i, modern Iranian -e, manuscript spelling, and editorial normalization are distinct layers of evidence.