How To Write a Sign Language, Part 4: History, Grammar & Glossing
Blog Post version available here: How To Write a Sign Language, Part 4: History, Grammar & Glossing
I'm back again! This is a bit of a different one. I am done with explaining the different types of sign language writing system and I want to discuss the uses. And first on the docket is an important question can these writing systems accurately reflect sign language grammar?
//
Whether you are new to sign languages or a native signer who has signed your entire life - you probably haven't heard about sign language writing systems. Spoken languages around the world have writing systems so surely sign languages can too? Well, many have been tried and used in brief corners of the world but none have been widely adopted by Deaf communities.
I recommend these resources if you want to learn more about all of those attempts and their histories:
- Sign Language Dictionary - Writing History
- The Hunt for The Perfect Sign Language Writing System (© 2025 Zrajm)
But in this mini-series of posts I want to explain the different types there are so that you, intrepid future sign language writer, may better understand the options and pick which to learn or make.
In the last part I talked about parameter alphabets, which tease out each individual part of a sign like a recipe. In this part I will talk about the grammar of sign languages and glossing!
Part 1: Logographies (Reddit Post & Comments Section)
Part 2: Pojectional Systems (Reddit Post & Comments Section)
Part 3: Parameter Alphabets (Reddit Post & Comments Section)
Part 4: History, Grammar & Glossing <= You are here
A Brief Tour of Deaf History
Before I begin we need a very brief history lesson.
Sign languages have always been around, but rarely is their history written down as Deaf people were mostly poor and illiterate - as were most people outside of the elite of society. Sporadic accounts tell of those who use their hands to talk from across the world back to the dawn of civilisation. Socrates even mentioned communication by signing amongst deaf people, albeit briefly. Sign languages spread and diverged for the same reasons that spoken languages do - people move to new places and then can't communicate with those back home where they came from, thus a new language is born.
But the oppression of deaf and hard of hearing people goes back as far also - as we are seen as stupid due to our inability to communicate, and attempts to "normalise" us are seen as virtuous.
Deaf Education began in 1760 with the founding of the Institution Nationale des Sourds-Muets à Paris by a man named Charles-Michel de l'Épée. There he, and other teachers both hearing and deaf, taught Deaf children using "manualism" (teaching via sign language). l'Épée learnt Old Parisian Sign Language (the sign language of the Deaf community of Paris) and tried to take those signs conform to the grammar of French - calling this signes méthodiques (methodical signs). This experiment quickly failed as the Deaf children and alumni of his school used their own grammar, becoming langue des signes française (LSF, aka French Sign Language). LSF would later be carried across the Atlantic by two men named Gallaudet and Clerc to become ASL.
Image Source: Charles-Michel de l'Épée - Wikipedia
In Milan 1880 the Second International Congress for Second International Congress on Education of the Deaf banned the used of manualism in favour of oralism - teaching deaf children only how to speak in the theory that it would make them more normal and better able to integrate into mainstream society. While this experiment worked for some, for many others it left them without accessible communication - and even for those who could integrate it often lead to difficulties socialising as they would forever struggle to understand those around them.
Example of Oralism: via manipulating the throat of a deaf child
Image Source: deaffriendly | flashBACK: Oral Method in Deaf Institutes (Post-Milan 1880)
Oralism as a practice continues in the modern day under new names and continues to do harm - the current recommended method by experts and researchers is the bilingual-bicultural model (BI-BI), teaching both sign and speech separately to bolster one-another.
During this period sign languages were not recognised as unique languages. Where they were used, they were often seen as a form of defective "slang". "ASL" and "BSL", while definitely in existence, had no names but instead were seen as broken English made with the hands.
And finally we get to what this story is about - Stokoe. William Stokoe is was a an English teacher at Gallaudet University - a university that specifically teaches Deaf students. As an English teacher and linguist he noticed his students signing and as he watched he began to realise their signs had consistent patterns behind them. They had set "phonemes (parameters) which he called "cheremes" (the name didn't stick), and grammar.
Image Source: Deaf History - Europe - 1960: William Stokoe, "Sign Language Structure"
This is what lead him to develop the Stokoe method, as well as coin the term "American Sign Language". That name did stick and soon sign languages around the globe had their own unique names and communities willing to fight for them.
Sign Language Grammar - The Big Prize
Sign languages have their own unique grammars that are different from anything a spoken language can produce. While much is the same, much is also different. And between sign languages there are many similarities and differences also.
Many basics are the same. Sign languages have nouns, verbs and adjectives. They have proper nouns (names) and pronouns. They arrange these into sentences to create meaning.
Sign language word order can also be starkly different. Many, if not most, sign languages work off a topic-comment system. This is usually where the topic (the item with more pre-existing information) comes before the comment (the item with more newly introduced information).
Consider the city of Paris. You already know what Paris is, and where it is. Thus it is more likely to be the topic, though could sometimes be the comment. If I was telling you that I did something in Paris - the new information is the information about what I did. You know who I am, you know what Paris is, but what I did there is new information to you. Thus topic comment usually produces something like the following word order:
- Time - when did it happen?
- Place - where did it happen?
- Object - who/what was involved?
- Subject - who/what did the things?
- Verb - what did they do?
- Question - what more information do I need to know?
I sometimes like to think of it like a theatre play. Think visually. Imaging the space in front of you is a stage:
Image Reference: Male and female artists playing on stage in front of audience. cartoon performers in evening dresses singing drama song flat illustration | Free Vector
- Time - first you tell the audience when they are
- Place - next you raise the curtain to reveal your backdrop telling them where they are
- Object - now you introduce the props your characters will use
- Subject - now your characters come onstage
- Verb - now your characters do stuff
- Questions - now you set up the hooks for the next scene
Not all sentences follow this, it is flexible. It can also produce Subject-Object-Verb and Subject-Verb-Object sentences. And different sign languages do this differently - some prefer OSV, some SVO some SOV. This isn't the be-all-end-all, just an overview.
Sign language pronouns are also unique and incredibly cool. You use a finger, or an entire hand, to point in a specific direction. Pointing towards oneself is ME, at the audience is YOU, at a random person is THEM-OVER-THERE, but from here it gets even cooler. Pointing at a random decided location in your personal space (known as "sign-space") sets that location up for later. So if I were to sign DOG THAT(left) then the left area of my signspace becomes the dog's area. Every time I point back in that direction, it refers to that dog. Additionally pointing in the same direction with a different handshape can be possessive (a fist ✊ in BSL, an open palm ✋ in ASL). Different sign languages point with different handshapes, and sometimes different parts of the body (eye gaze, nose, tongue, lips). This is all called "indexing".
Image Source: Indexing American Sign Language (ASL)
Verbs are also very often directional. As I showed in Part 2, verbs can move in different directions depending on who is doing what to what.
Image Source: American Sign Language (ASL)
But this also often agrees with the indexing - where verbs will move towards and way from these placed indexes. If I pet that dog from the previous example, I would sign STROKE towards the direction that the dog is.
Roleshifts are another uniquely sign language part of grammar. These are where you shift your shoulders, along with your demeanour, to pretend to be another person. These are often used to quote others. Rarely to signers say "She said..." or any equivalent, they just role-shift and become the person saying the thing for a small while.
Classifiers are also a big topic. These are also known as "pro-forms" or "depictive signs". Essentially classifiers are "show-don't-tell" within a language. I like to break it down into three or four categories.
- Entity Classifiers - with these your entire hand becomes the thing, and like a puppet you can move it about through the air to imitate it's movement. For example the 🤙 handshape in BSL can be an aeroplane, and you can fly it about through the air steadily before shaking violently to show the horrendous turbulence in your last flight.
- Handling Classifiers - with these you just pretend to hold and use objects. For instance you could hold a bowl and a whisk and mix it about furiously to show how hard it is to make meringue.
- Bodily Classifiers - with these you become the thing you are talking about. For instance you could pretend to be a cat and pounce on a mouse.
- Tracing Classifiers - with these you trace out the shape of something. For instance you could trace out a line of vases that are different sizes and shapes - one square, one round, one voluptuous.
Entity Classifiers in ASL: ☝️ = person, ✌️ = legs, 👆+🖕 = vehicle
Image Source: Classifiers: a list of CL handshapes
All Deaf Community Sign Languages (those used by a Deaf community - where the majority of signers are Deaf) use classifiers so far as I am aware. This is, however, not true of sign languages outside of those such as Village Sign Languages, where the majority are hearing with a higher than usual amount of deaf individuals.
But even amongst Deaf Community Sign Languages - there is plenty of range of what classifiers are used and how. For instance while the 🤙 is a plane in BSL, it is a person in some far east Asian sign languages (I believe one of the Chinese Sign Languages at least). And in America, airplane is 🤟. Also the classifier usually goes after a verb but before a question in the topic-comment structure listed above.
Lastly I want to very briefly touch on expressions (NMFs) - as these can do a lot of work with the grammar. Not all of this is done in a sign. The tilt of a shoulder or change of expression can signal something that you need to know in order to make sense of what is being signed.
Sign languages were also heavily influenced by spoken languages - both by active attempts to make sign languages follow spoken language rules such as l'Épée's attempt back in 1760 or modern ones such as Signed Exact English - and also by bleed over from one language to another. As such three layers exist:
- Sign Languages - full sign languages with sign language grammar, such as ASL, BSL, LSF etc.
- Contact Languages - such as Pidgin Signed English (PSE) (also known as Conceptually Accurate Signed English - CASE) which mix ASL and English grammar.
- Manually Coded Languages - sign systems which use signs following the word order of spoken languages, such as Signed Exact English (SEE).
The boundaries between each of these is often fuzzy, with a small number of signers signing full sign languages with not a hint of spoken language grammar and few signing full manually coded languages with not a hint of sign language grammar - but the majority of signers hovering somewhere between. Where to short-cuts become a contact language? Where do some influences of English on one's signing become a contact language? These are questions that those who wish to preserve sign languages have to wrestle with.
As you can hopefully see, sign language grammar functions on principles that are adjacent to, but different from spoken languages. This is, of course, just an overview - and to learn more you should learn your local sign language. Understanding and representing it correctly is the big prize amongst both linguists and those who wish to write sign languages.
GLOSS
But through all the parts of this so far, some of you may have noticed a glaring absence, the most widely used sign language writing system of all - GLOSS!
Glossing isn't unique to sign languages, it is a tool used by linguists to analyse grammar. Outside sign languages it's usually called interlinear gloss or Leipzig glossing and matches the words and morphemes of one word with words and morphemes with another next to one another:
That is an easier to read example, but often it looks something like this:
Image Source: Interlinear gloss - Wikipedia
The small caps word such as "1SG" and "DET" mark the grammar of the language.
- 1SG - first person singular
- SUBJ - subject
- 3SG - 3rd person singular
- OBJ - object
- APPL - applicative voice
- DET - determiner
- POSS - possessive
The punctuation has a specific usage, usually a dash means separate morpheme (meaningful part of a word) whereas dot means part of the same morpheme.
Words that can be translated into the "metalanguage" (the one being used to do the analysis) are given in lower case.
The rules of sign language glossing. First off, as sign languages don't have accepted writing systems, they are more commonly presented on their own - though sometimes alongside videos, images or drawings of signing. e.g.
- DOG
- CAT
Second, sign language glosses are majority written in capital letters, with hyphens between words when they refer to a singular sign / morpheme. e.g.
- STARE-AT
- GO-TO
Third, they have various strategies for representing sign language grammar:
- "SIGN+SIGN" - compound signs
- "SIGN+", "SIGN++", "SIGN+++" etc - repeated / reduplicated signs
- "fs-WORD" or "FS-WORD" or "(fs)WORD" etc - fingerspelling
- "#WORD" - fingerspelling loan signs, like #DOG in ASL.
- "IX1", "IX2", "IX-left", "IX-she" etc - indexing, aka pronouns,
- "CL:1 (walk slowly)", "CL:V (running fast)", "CL:3 (zoom past)" etc - classifiers, "CL" just means "classifier", whereas 1, V, 3, etc shows the handshape and the part in brackets tells the reader what action occurred with the classifier, usually in a short English phrase or snippet
- "/\" - eyebrows raised (rare)
- "\/" - eyebrows lowered (rare)
- Other information such as NMFs can usually go in brackets in lower case.
For more on gloss I recommend:
- ForsterJensSteinDanielOrmelEllenCrasbornOnnoNeyHermann--BestPracticeforSignLanguageDataCollectionsRegardingtheNeedsofData-DrivenRecognitionTranslation--2010.pdf
- Gloss in American Sign Language (ASL)
- Glossing: transcription symbols in sign language
- Rules-of-Glossing-in-ASL.pdf
I have most widely seen gloss used in America and sporadically elsewhere in the world. I have heard of it being taught in sign language classes there which is unheard of here in Britain. It is likely the widest used sign language "writing system", especially in America, but there are flaws to that. Namely the fact that the words aren't actually in ASL, they are English words written with the grammar of ASL, thus mismatches occur.
There are precedents for languages using words from others to write their own.
Image Source: Pahlavi scripts - Wikipedia
The Pahlavi script was a script used to write various middle Perisian and other nearby related middle Iranian languages. It was derived from Aramaic and used "heterograms", which were spellings which didn't match the current word. These heterograms were from Imperial Aramaic, a completely different language unrelated to Persian.
But even the Pahlavi script used a mixture of those Persian words in it. The heterograms were there to make it more understandable across audiences, but those words necessary for understanding (especially the grammar) were in the language(s) the people actually used.
'For example, the word for "dog" was written as ⟨KLBʾ⟩ (Aramaic kalbā) but pronounced sag; and the word for "bread" would be written as Aramaic ⟨LḤMʾ⟩ (laḥmā) but understood as the sign for Iranian nān. [...] [these] could also be followed by letters expressing parts of the Persian word phonetically, e.g. ⟨ʾB-tr⟩ for pidar "father". The grammatical endings were usually written phonetically.'
Source: Pahlavi scripts - Wikipedia
Just the heterograms alone would be insufficient to accurately write Persian, and the writing system was eventually dropped in favour of unambiguous alternatives such as the Avestan Alphabet and later the Arabic abjad.
Similarly Gloss is useful, but only half way there. It can express the basic word order of ASL decently well, and denote when certain grammatical features occur - but it often creates situations where a single word in gloss could be multiple signs, and for classifiers it requires writing in snippets of English or whole English sentences. This means that if you are a Deaf child, you have to learn how to write English (a whole different language) before you can write your own).
I say this as someone who has tried. I have written two stories that use gloss to represent the signing of various characters... and while interesting, it was quite insufficient in my experience.
Example of Glossing Used for Dialogue
The above is an extract from the English translation of my short story Y Ddwyleg, published in Welsh in the Gwyllion Magazine: Gwyllion issue 9 Digital Download – Gwyllion.
S. Leigh Ann Cowan who wrote Ranking Deaf Characters in Fiction: Reviews, Analyses, and Tips to Write Better Deaf Characters (Updated Monthly) shared with me an upcoming guide to writing Deaf / signing characters in fiction which she says "I don’t believe glossing is appropriate in literature."*
*this is quoted with her direct consent and may not reflect the wording present in the final released document.
She notes that most examples of gloss in writing are simplified (thus missing many of the grammatical features), ambiguous and appear as broken English, which is especially noticeable when paralleled with characters using fully fluent spoken English. Instead she recommends writing in fully fluent translated sentences for characters in books*. I will talk more about this in my next part covering when and whether it is appropriate (or not appropriate) to use any form of sign language writing - be that gloss or any other.
*her review of my work with gloss in it was positive, despite being against her general advice, but only because I strongly understood what I was doing and my work was very experimental
As such Deaf people do not tend to use gloss when writing much, especially not for anything longer than a few words. This might be partially because gloss often looks like "broken English" - but also it loses the visual part of sign language. A signed sentence feels full and alive, written down on the page as gloss it feels empty and dead, dissected using gloss.
Sign Language Grammar in Sign Language Writing
So unlike gloss, which is writing one language with the words of another, sign language writing has the opportunity to tap directly into the languages themselves.
In the part on Parameter Alphabets I discussed how writing systems can be used to break down the phonology of sign languages. Some only aim to do that such as HamNoSys and explicitly do not aim to be written in longer sentences. But others do.
This presents somewhat of a problem because the current norms for writing sentences and longer texts are based off the norms of spoken languages. Full stops, commas, colons, exclamation marks, question marks. How much of these can be taken wholesale? How much needs to be adapted for sign languages? How much new stuff needs to be invented?
I find that if you just go into this writing word by word, sign by sign, something tends to be lost. Across multiple signs information is conserved, and done so in a very visual way. For instance with the indexing and verb directions. Planning ahead so it all works well together is important.
Sign Language Grammar in Logographies
There aren't really enough logographies to say if there is a pattern in how they handle it in general - but they can either handle it by depicting what is going on more directly - or introduce symbols that indicate specific grammatical things are going on.
from a previous blog post
While the grammar of Handtalk Pictographs is unclear from the materials I have found - you can observe some modification and arrangement of symbols that seem to indicate... something. What precisely that is, I am unsure. I believe repetition can indicate number via reduplication. There are also modification of the "tracks" / "footsteps" glyph in various situations.
Image made with: Word-by-word translation visualizer — see which words match · Bitext Align
Icoglyphs uses many tricks to accurately capture sign langauge grammar as best as possible. One such trick is glyphs specifically meant to indicate grammatical location. Specifically in the example above:
These are grammatical markers that inform the reader of the locations and direction of various signs.
It also has ways of making classifiers using classifier brackets and arrows:
I am biased towards logographies - so I think they do a pretty good job at representing sign language grammar. However they require a lot of memorisation.
Sign Language Grammar in Projectional Systems
As I explained and demonstrated in Part 2, directionality is one key benefit of projectional systems. You can modify words in order to visually incorporate directionality very easily!
Classifiers are also possible, though I feel sometimes without a logographic element they don't quite feel like classifiers. It's hard to place my finger on precisely but again they feel like a diagram rather than a show-don't-tell. In fact it feels like telling me how to show.
Jack and Jill in Sutton Signwriting
In the above example, most of the words are classifiers. The translation would be something like:
- Person one and person one got together - entity classifiers
- Went up together - entity classifiers
- THINK (not a classifier)
- Bucket - tracing classifiers
- Fill up - entity / tracing classifiers
- Person one tumble down - entity classifiers
- Head hurt - bodily classifiers
- Person two tumble down - entity classifiers
"Two people got together and walked up a hill. They thought to fill two buckets. One fell down. The other does too."
It proves it is possible, but I first have to decode this into signs before I can understand it as classifiers.
Sign Language Grammar in Parameter Alphabets
I won't cover this in all alphabets but luckily I don't have to. Many alphabets don't have many adaptations specifically for sign language grammar, instead they write word by word and hope that is good enough.
Signfont uses expression (NMF) markers as punctuation, mostly before the rest of the sentence:
As far as I am aware the rest of the system makes no special cases for directionality nor classifiers.
ASLfont shows indexing and sign-space locations in a very clear way that could help with sign language grammar recognition:
Image Source: ASL Font: Lesson 6: Locations in Space 1
Location is also marked in a very distinctive way in ASLwrite, with -T and -J being which side of the signspace and entire clause is located:
>
Lentz et al. 1993, 45.
These are especially useful for role-shifts. It also has systems for swapping hands and using buoys, both of which are concepts I haven't talked about but are also important grammatical concepts.
Overall - parameter alphabets can and should make unique adaptations for sign language grammar - but each one makes completely different adaptations. This means there is a lack of consistency across them.
This is, of course, the big prize I think a good sign language writing system could win - a way to express not only the phonology (parameters) of sign languages but also the grammar in an unambiguous, clear and efficient way. I think this would also help Deaf signers and learners of sign languages feel more confident about their grammar - as they can see it written down and build stronger associations. It may also help to differentiate sign languages from contact languages or manually coded languages.
Conclusion
Gloss currently exists as a way to represent sign language grammar - but fails as a writing system because it is unwieldy, often ambiguous and requires you to master English (a different language) before being able to write your own language.
Thus sign language writing systems have a huge opportunity to express sign language grammar well, and I believe must in order to succeed. It is the big prize, because writing a single word is "easy", taking the next step to writing not only a single short sentence but longer complicated ones and whole paragraphs, chapters and books needs skill and mastery of all of a language's grammar.
I recommend that anyone looking to make a sign language writing system to think ahead for grammar and not neglect it as you are making the system. I also recommend you dedicate a significant portion of your learning/teaching materials to addressing the grammar.
I recommend those who are considering choosing a writing system to learn and use look at how it represents grammar before making your decision.
I recommend writers aiming to write in sign language think about sign language grammar rather than just write word by word. Learn how to structure your overall sentences, punctuation, paragraphs and formatting to do so most effectively in your writing system of choice.
