{{Short description|Non-printing character that separates two normally joined characters}} thumb|upright|ISO keyboard symbol for ZWNJ|class=skin-invert-image thumb|upright|A ZWNJ between the double-wide tilde and the acute accent centers the acute over the tilde, instead of over the ''eͤ'', where it would appear otherwise.<ref>Navarro Tomás (1962) Atlas lingüístico de la la Península Ibérica (ALPI), tomo 1 ‘Fonética.’ Consejo superior de investigaciones científicas, Madrid. From map 69, location 240.</ref>|class=skin-invert-image The '''zero-width non-joiner''' ('''ZWNJ''', rendered: {{Not a typo|{{kbd|‌}}}}; HTML entity: {{kbd|&zwnj;}} or {{kbd|&#8204;}}) is a non-printing character used in the computerization of writing systems that make use of ligatures. For example, in writing systems that feature initial, medial and final letter-forms, such as the Persian alphabet, when a ZWNJ is placed between two characters that would otherwise be joined into a ligature, it instead prevents the ligature and causes them to be printed in their final and initial forms, respectively. This is also an effect of a space character, but a ZWNJ is used when it is desirable to keep the characters closer together or to connect a word with its morpheme.
The ZWNJ is encoded in Unicode as {{unichar|200C|ZERO WIDTH NON-JOINER|html=}}.
== Use of ZWNJ for correct typography == In certain languages, the ZWNJ is necessary for unambiguously specifying the correct typographic form of a character sequence.
{|class="wikitable" style="font-variant-ligatures: common-ligatures discretionary-ligatures;" !colspan="3"| Correct (with ZWNJ) !colspan="3"| Incorrect !rowspan="2"| Meaning |- ! Display ! Picture ! Code ! Display ! Picture ! Code |- | {{script/Arabic|size=200%|attributes=lang="fa"|میخواهم}} | class=skin-invert-image | <code>می&zwnj;خواهم</code><br /><br /><small>''(rendered from right to left):''<br /></small><div dir="rtl"><code>می<span dir="ltr">&zwnj;</span>خواهم</code></div> | {{script/Arabic|size=200%|attributes=lang="fa"|میخواهم}} | class=skin-invert-image | <code>میخواهم</code> | Persian 'I want to' |- | {{script/Arabic|size=200%|attributes=lang="ms"|سا{{Jawi-HamzaThreeQuarter|8}}ينس}} | 74px|class=skin-invert-image | <code>ساءين&zwnj;س</code><br /><br /><small>''(rendered from right to left):''<br /></small><div dir="rtl"><code>ساءين<span dir="ltr">&zwnj;</span>س</code></div> | {{script/Arabic|size=200%|attributes=lang="fa"|سا{{Jawi-HamzaThreeQuarter|8}}ينس}} | 74px|class=skin-invert-image | <code>ساءينس</code> | Malay 'science' |- | {{script/Arabic|size=200%|attributes=lang="ks"|ۂسؠتِنؠ}} | | <code>ۂسؠ&zwnj;تِنؠ</code><br /><br /><small>''(rendered from right to left):''<br /></small><div dir="rtl"><code>ۂسؠ<span dir="ltr">&zwnj;</span>تِنؠ</code></div> | {{script/Arabic|size=200%|attributes=lang="ks"|ۂسؠتِنؠ}} | | <code>ۂسؠتِنؠ</code> | Kashmiri 'A female elephant' |- | {{Script/Hebrew|size=150%|attributes=lang="hbo"|הֱֽיֹות}} | class=skin-invert-image | <code>הֱ&zwnj;ֽיֹות</code><br /><br /><small>''(rendered from right to left):''<br /></small><div dir="rtl"><code>הֱ<span dir="ltr">&zwnj;</span>ֽיֹות</code></div> | {{Script/Hebrew|size=150%|attributes=lang="hbo"|הֱֽיֹות}} | class=skin-invert-image | <code>הֱֽיֹות</code> | Old Hebrew 'be' |- | {{lang|de|2=<span style="font-family:serif; font-size:x-large;">Auflage</span>|italic=no}} | class=skin-invert-image | <code>Auf&zwnj;lage</code> | {{lang|de|2=<span style="font-family:serif; font-size:x-large;">Auflage</span>|italic=no}} | class=skin-invert-image | <code>Auflage</code> | German 'edition' (compound of "{{lang|de|auf}}"+"{{lang|de|Lage}}") |- | {{script|de-Latf|2=<span style="font-size:x-large;">Brot‌zeit</span>}} | class=skin-invert-image | <code>Brot&zwnj;zeit</code> | {{script|de-Latf|2=<span style="font-size:x-large;">Brotzeit</span>}} | class=skin-invert-image | <code>Brotzeit</code> | German (regional) '(kind of) snack'<br />(compound noun "{{lang|de|Brot}}"+"{{lang|de|Zeit}}" = 'bread time'),<br /> shown in Fraktur |- | <span style="font-family:serif; font-size:x-large;">deafly</span> | class=skin-invert-image | <code>deaf&zwnj;ly</code> | <span style="font-family:serif; font-size:x-large;">deafly</span> | class=skin-invert-image | <code>deafly</code> | Not a compound of "dea"+"fly", but the adverb of "deaf" |- | {{lang|ne|2=<span style="font-family:serif; font-size:x-large">श्रीमान्को</span>}} | | <code>श्रीमान्&zwnj;को</code> | {{lang|ne|2=<span style="font-family:serif; font-size:x-large">श्रीमान्को</span>}} | | <code>श्रीमान्को</code> | In Nepali "of husband" or "of respected person" according as what "श्रीमान्" is used to represent (husband or respected person) |- | {{lang|bn|2=<span style="font-family:serif; font-size:x-large">উদ্যাপন</span>}} |164px|class=skin-invert-image | <code>উদ্&zwnj;যাপন</code> | {{lang|bn|2=<span style="font-family:serif; font-size:x-large">উদ্যাপন</span>}} |148px|class=skin-invert-imagex | <code>উদ্যাপন</code> |Bengali {{gloss|celebration}} |- | {{lang|ne|2=<span style="font-family:serif; font-size:x-large">अय्लाः</span>}} | | <code>अय्&zwnj;लाः</code> | {{lang|ne|2=<span style="font-family:serif; font-size:x-large">अय्लाः</span>}} | | <code>अय्लाः</code> | {{lang|new-Latn|Aila}}, a traditional alcoholic beverage, in Nepal Bhasa |- | {{lang|te|2=<span style="font-family:serif; font-size:x-large">హైద్రాబాదు</span>}} | 123px|class=skin-invert-image | <code>హైద్&zwnj;రాబాదు</code> | {{lang|te|2=<span style="font-family:serif; font-size:x-large">హైద్రాబాదు</span>}} | | <code>హైద్రాబాదు</code> | ''Hyderabad'' in Telugu |} The picture shows how the code looks when it is ''rendered'' correctly, and in every row the correct and incorrect pictures should be different. On a system which is not configured to display the Unicode correctly, the correct display and the incorrect one may look the same, or either of them may be significantly different from the corresponding picture.
In this Biblical Hebrew example, the placement of the {{lang|he-Latn|meteg}} to the left of the {{lang|he-Latn|segol}} is correct, which has a {{lang|he-Latn|shva}} sign written as two vertical dots to denote short vowel. If a {{lang|he-Latn|meteg}} were placed to the left of {{lang|he-Latn|shva}}, it would be erroneous. In Modern Hebrew, there is no reason to use the {{lang|he-Latn|meteg}} for spoken language, so it is rarely used in Modern Hebrew typesetting.
In German typography, ligatures may not cross the constituent boundaries within compounds. Thus, in the first German example, the prefix {{lang|de|Auf-}} is separated from the rest of the word to prohibit the ligature ''fl''. Similarly, in English, some argue ligatures should not cross morpheme boundaries.<ref>{{cite web |url=https://english.stackexchange.com/questions/50660/when-should-i-not-use-a-ligature-in-english-typesetting#answer-50957 |title=When should I not use a ligature in English typesetting? |work=english.stackexchange.org }}</ref>{{Better source needed|date=November 2018}} For example, in some words ''fly'' and ''fish'' are morphemes but in others they're not; therefore, by their reasoning, words like ''{{not a typo|deaf‌ly}}'' and ''{{not a typo|self‌ish}}'' (here shown with the non-joiner) should not have ligatures (respectively of fl and fi) while ''dayfly'' and ''catfish'' should have them.
Persian uses this character extensively for certain prefixes, suffixes and compound words.<ref>{{cite web |url=http://persian.nmelrc.org/persianword/zwnj.htm |title=The Zero-Width-Non-Joiner |work=National Middle East Language Resource Center|archive-url=https://web.archive.org/web/20120708235820/http://persian.nmelrc.org/persianword/zwnj.htm |archive-date=July 8, 2012}}</ref> It is necessary for disambiguating compounds from non-compound words, which use a full space.
In the Jawi script of Malay, ZWNJ is used whenever more than one consonants are written at the end of any phrase ({{lang|ms-Arab|سا{{Jawi-HamzaThreeQuarter}}ينس}}, Malay for 'science' or {{lang|ms|sains}} in Latin script, pronounced /ˈsa.ɪ'''ns'''/.) It is used to signify that there are no vowels (specifically 'a' or 'ə') in between the two consonant letters as {{lang|ms-Arab|سا{{Jawi-HamzaThreeQuarter}}ينس}} would otherwise be pronounced either /ˈsa.ɪnas/ or /ˈsa.ɪnəs/. A space would separate the phrase into different words, where phrases such as {{lang|ms-Arab|سا{{Jawi-HamzaThreeQuarter}}ين س}} would now mean 'to sign the Arabic letter sin' ({{lang|ms|sain sin}} in Latin script.)
In Kashmiri written in the modified Perso-Arabic script, the behavior of ؠ (Kashmiri Yeh) requires specific orthographic handling in digital text to maintain both visual and phonetic accuracy. This character indicates palatalization and possesses two principal typographic forms: the '''Gol Ye''' (with a diacritic ring) and the '''Taler Ye''' (a truncated form without the ring). Standard rendering protocols typically display the Taler form in word-final and isolated positions. However, these two forms are phonetically distinct and are often treated in pedagogical contexts as separate characters. The Taler Ye represents the palatalization of the preceding consonant {{IPA|[◌ʲ]}}, whereas the Gol Ye in the medial position represents the palatalized vowel sound {{IPA|[ʲa]}}. In certain lexical contexts, such as {{lang|ks|شُپؠیَن}} (Shopian), a medial Taler form is linguistically mandatory to preserve the correct pronunciation and spelling. Because standard digital shaping engines often default to the joined Gol form in initial and medial positions, the '''Zero Width Non-Joiner''' (ZWNJ) must be used to suppress this joining behavior. This ensures the character retains its correct Taler shape within a single word unit, fulfilling the precise orthographic conventions of the script.<ref name="Unicode2024">{{cite web |last1=Goregaokar |first1=Manish |last2=Pournader |first2=Roozbeh |title=Fixing the Kashmiri Yeh |url=https://www.unicode.org/L2/L2024/24152-kashmiri-yeh.pdf |publisher=Unicode Technical Committee |date=2024-06-23 |access-date=2026-05-03}}</ref>
== Use of ZWNJ to display alternative forms == [[File:Zero Width (Non)Joiner.svg|thumb|Use of ZWNJ and {{control code link|ZWJ}} to select alternative forms of Devanagari, Tamil, Kannada, Sinhala and emoji.]] In Indic scripts, insertion of a ZWNJ after a consonant either with a halant or before a dependent vowel prevents the characters from being joined properly:<ref>{{Cite web|url=https://www.unicode.org/faq/indic.html#17|title=FAQ - Indic Scripts and Languages|website=www.unicode.org|access-date=2020-03-15}}</ref>
In Devanagari, the characters {{script|Deva|क्}} and {{script|Deva|ष}} typically combine to form {{script|Deva|क्ष}}, but when a ZWNJ is inserted between them, {{script|Deva|क्ष}} (code: <code>क्&zwnj;ष</code>) is seen instead.
In Kannada, the characters ನ್ and ನ combine to form ನ್ನ, but when a ZWNJ is inserted between them, ನ್ನ is displayed. That style is typically used to write foreign words in Kannada script: "Facebook" is written as ಫೇಸ್ಬುಕ್, though it can be written as ಫೇಸ್ಬುಕ್. ರಾಜ್ಕುಮಾರ್ and ರಾಮ್ಗೊಪಾಲ್ are examples of other proper nouns that need ZWNJ.
To insert a ZWNJ in Kannada, use Shift-V on Linux (iBus, InScript). On Windows (InScript), you can produce a ZWNJ with Ctrl+Shift+2 or Alt+0157. For the LipikaIME on Mac, the caret returns a ZWNJ.
In Bengali, when the Bengali letter য occurs at the end of a consonant cluster—i.e., য preceded by a ◌্ (''hôsôntô'')—it appears in a special shape, alt=Rendering of Bengali Ja-phala|frameless|x15px, known as the য-ফলা (''ja-phala''), such as in ক্য (ক ্ য). Thus, when we want to write উদ্যাপন (correct Bengali spelling for celebration), it becomes উদ্যাপন (which is incorrect). Here ZWNJ works. If we want to write উদ্যাপন, we have to write in the following sequence (code: <code>উদ্&zwnj;যাপন</code>),<ref name=":0">{{Cite web |title=Bengali FAQ in Unicode. |url=https://www.unicode.org/faq/bengali.html#6}}</ref><ref name=":1">Also see the Unicode chapter 12, Bengali (Bangla) between page 475 to 479 here in [https://www.unicode.org/versions/latest/ch12.pdf#G664195 PDF].</ref> then we will get the proper rendering and the correct spelling. In Bengali, the ''hôsôntô'' is used for making any conjuncts and falas (such as ra-fala, ba-fala etc)''.'' Where the ''hôsôntô'' needs to be displayed explicitly, it is required to insert ZWNJ after the ''hôsôntô''.
Also in Bengali, when the Bengali letter র occurs at the beginning of a consonant cluster—i.e., র succeeded by a hôsôntô—it appears in a special shape, known as the রেফ (reph). Thus, the sequence র ্ য is rendered by default as র্য. When the য-ফলা shape needs to be retained rather than the রেফ shape, the ZWJ {{unichar|200D|ZERO WIDTH JOINER|html=}} is inserted right after র, i.e., র&zwj;্য to render র্য.<ref name=":0" /><ref name=":1" /> র্য is commonly used for loanwords from English such as র্যাম (RAM), র্যান্ডম (random) etc.
== Symbol == [[File:German-T2-Keyboard-Prototype-May-2012-part.jpg|thumb|German T2 keyboard (detail), showing the ZWNJ symbol on the "." key]] The symbol to be used on keyboards which enable the input of the ZWNJ directly is standardized in Amendment 1 (2012) of ISO/IEC 9995-7:2009 ''"Information technology – Keyboard layouts for text and office systems – Symbols used to represent functions"'' as symbol number 81, and in IEC 60417 ''"Graphical Symbols for use on Equipment"'' as symbol no. IEC 60417-6177-2.
== Usage in Microsoft Word == In Microsoft Word's special-character insertion function, the zero-width non-joiner is called a "No-Width Optional Break".<!-- Found via Insert -> Symbol -> More Symbols -> Special Characters. Some sources seem to say this is 200B in Word, but it's easy to show that Office365 is currently inserting 200C. It may be version-dependent. -->
== See also == * Zero-width joiner * Zero-width space * Word divider
== References == {{Reflist}}
== External links == * [https://web.archive.org/web/20190219182722/http://sartre2.byu.edu/persian/persianword/zwnj.htm Using the ZWNJ in Persian] * [http://archives.miloush.net/michkap/archive/2006/02/15/532394.html Every character has a story #19: U+200c and U+200d (ZERO WIDTH <nowiki>[NON]</nowiki> JOINER)]
{{Unicode navigation}}
Category:Control characters Category:Persian orthography Category:Typography Category:Unicode formatting code points