What's a Word Worth? - The Trichotomy Tango (Peircian Semiotics #2)

[justira]

What's a Word Worth is a new column by justira about the mechanics of writing. In this column, I examine the actual writing of every single book I read, focusing on how it conveys meaning and whether the writing works for me as an editor, reader, and fellow writer. My analysis will be based on the Peircian semiotic framework, explained in the first few posts of the column.

Welcome to the second post in my introductory series for this column! The first four posts will introduce readers to Peircian semiotics, which is the framework I use to analyze writing. Please check out the first post in this series, especially the review of terms.

(And before we get going, yes, I know I named a post about trichotomies after a quadruple meter dance, but I couldn't resist the alliteration, okay? >.>)

Let's jump right in to the first⁽¹⁾ trichotomy of the sign-relation!

Trichotomy I: Types of Representamena

This first trichotomy of the sign-relation involves the nature of the representamen, and is one of Peirce's contributions that has survived somewhat intact into the general literature of linguistics in the form of the type-token distinction⁽²⁾. Peirce himself gave these categories various names over the course of his writing and development, but I'm gonna use the most common names along with their more vernacular equivalents: qualisign (tone), sinsign (token), and legisign (type). I'm going to briefly go over definitions first, then give examples. I'm also going to use a lot of examples from phonology, since that's such a good example of this trichotomy, but I'll go over or link to terminology for the unfamiliar.

A qualisign or tone is a sign that is pure quality — an abstract, distinct characteristic or feature. Examples of qualisigns are the colour of an object or the pitch of a musical note. Linguists are most familiar with qualisigns in the context of phonetics, where they are qualities such as "voiced" (in the phonetic sense) or "fricative". Qualisigns are prime examples of Firstness.

A sinsign or token is an actual occurrence or fact; it is a singular actually existent object in the world, a specific instance of a sign. A specific red object, like a red rose, is a sinsign with the qualisign of redness. A note someone actually sings or plays is a sinsign with the qualisign of its pitch.

I said earlier that qualisigns are Firstnesses. Firstnesses are embodied or instantiated — represented by an instance — in Secondnesses: qualisigns are instantiated in sinsigns. Sinsigns are a Secondness because they are actual objects or events and embody a relation between two entities. In the case of the sinsign, the relation is with its qualisigns: the sinsign embodies or instantiates those qualisigns. Note that sinsigns, to put it loosely, have to "actually happen": they have to be specific instances of things. As soon as you start talking about classes of objects — like "all red roses" or "all notes of this pitch" — we are moving from Secondness to Thirdness.

This Thirdness is the legisign or type: a law that is a sign; an abstract law of association independent of any particular instantiation of it. Now, legisigns are particularly hard to understand, so I'm going to go into an extended example below. For now, just remember that legisigns, being laws, are Thirdness: they involve a thinking being's mediational capabilities to form general relationships between two other things. Like when I say "all red roses", how do you know what I mean by "red"? Or "roses" for that matter? Not just the words, but the meaning of them: what are the "boundaries" of redness in roses? What is "red enough"?

Fig 2.1: Three roses.
People who use the English colour terms and can see colour can all agree that the top one is red (source). What about the one on the left? Is it red enough? (source) Is the one on the right red or orange or pink? (source)

The law-like relation that binds some items but not others under the aegis of "red rose" is an example of how legisigns work.

Now let's do one extended, specific example from phonology.

In linguistics, phonology provides us with great examples of each of the above types of representamen. First, a notational convention:

• [ ] Square brackets enclose specific occurrences of a phenomenon, such as a word. So, for example, if I say "dog" at 11:49 AM on Tuesday, September 20, 2016, then this is transcribed [dog]. This is an example of a token-level phenomenon.


• / / Slashes enclose type-level phenomena. So when I say [dog], this is an instance of the word, /dog/. /Dog/ is an idea, not attached to any particular speaker but existing as part of the language.

Thus, I might say [dog] and through its particular combination of phonetic properties (qualisigns) you can recognize is as being an instance (sinsign) of the word (legisign) /dog/.

So, phonology! Linguists are familiar with the concept that there is the idea, distinct from all other ideas, of a sound /t/ in English, and that this is different from the similar but distinct sound, /d/. Try saying "tip" and "dip" and pay attention to the position of your tongue. They're very similar, yes? But we consider them distinct sounds in the English language. Likewise, we take various qualitatively different sounds, such as the aspirated [t^h] in tip and the unaspirated [t] in stand to be instances of the same "idea" /t/. In linguistic terms, these are called phones (tokens) of the same phoneme (type). And the qualities of each occurrence [t] that we find relevant to making it an instance of /t/ — the voiced lingua-alveolar-plosiveness — are the qualisigns (tones) that determine its potential to serve as a sinsign of the legisign /t/.

Similarly, we have an idea for a word in English, /dog/, that is distinct from the ideas for the words /bog/ and /cat/ and /human/ and /chair/ and /ocean/. Each time I say "dog", it is an instance, [dog], of the legisign /dog/.

Notice that sinsigns are the only kinds of signs that are single, existing objects. Qualisigns and legisigns are both dependent on realization through and instantiation in sinsigns. Put simply, qualisigns and legisigns have to happen in the form of sinsigns. Legisigns must each have a multitude of sinsigns to support their categorical nature: a class- or type-level phenomenon can't be defined around a single object, but can only be defined around multiple objects that can be classed together. Likewise, any sinsign is dependent upon the qualisigns, the qualities of the sign, that allow us to apprehend it, and our understanding of a sinsign is informed by it being an instance of a particular legisign. This is how I can say [dog] and you recognize it, through its qualisigns, as a sinsign of the legisign /dog/.

Now imagine two instances of /dog/, like one I speak and one you speak. We might have slightly different pronunciations, or we could have different accents, but both will be recognized as tokens, [dog], of the type /dog/. Or, imagine I say "dog" and also write the word "dog". Even though these are completely different mediums, we still recognize them as being of the type /dog/. The fact that any two tokens [dog] can be identified as being "the same kind of thing" (even though the manifestation may be different in all kinds of ways) and are taken to refer to the same object attests to the reality of legisigns. Similarly, the colour red is a legisign, instantiated in such vastly different shades as the red of lips or the red of a rose or the red of a stoplight. Legisigns themselves — the classes by which we group phenomena — are generally nested within other such type-level classes, so that /cat/ as an idea is nested within classes like /pets/, /mammals/, or /animals to which many people are allergic/.

Legisigns have been recognized as socially relative categories by which we conceptually group phenomena: in other words, legisigns are a foundational aspect of culture⁽³⁾. Aliens might have totally different legisigns from us. In fact, if they don't, I would find that very odd. What is possibly less recognized is that qualisigns, too, are socially relative categories: what "counts" as a quality is socially determined. This idea is crucial to understanding the power of icons, which is what we'll be covering next. Our judgements about what counts as similarity are based on qualities we find important or interesting, what qualities count towards the semiotic potential of a sign; what, indeed, counts as a quality at all.

Aliens would not only have different legisigns, but they would almost certainly be built up of different qualisigns, perhaps qualisigns we don't or can't understand at all. A good example of this is the Presgr aliens from the Imperial Radch books. They have a category called Significance, which is a legisign. What makes a species Significant are the qualisigns of that legisign. Instances of Significant species — sinsigns of the legisign — are, well, a spoiler. But it's an interesting question, no?

Trichotomy II: Representamen-Object Relations

This second trichotomy is one of Peirce's better-known and less-understood contributions to linguistics. It deals with the ways a representamen may stand for its object: by being an icon, index, or symbol.

An icon, roughly speaking, denotes its object by sharing qualities with it. So, an icon refers to an object by being similar to it in some way. Note that an icon possesses these qualities, which give it the potential to act as a sign, regardless of the object and interpretant. Moreover, icons are the realm of possibility (Firstness), for a given icon cannot point to a specific object but only refers to the possibility of an object that shares qualities with it. So, a stick-figure drawing of a human is iconic of a human because, like a human, it has a head, torso, four limbs, and stands upright.

I'll go over the basis of such resemblance in greater detail later, but for now it's important to note two things. First, similarity is culturally grounded, as established above. The stick figure may not appear very representative of a human if an alien were looking at it. Or if the alien shares those qualities with humans, the alien might think it's a representation of their species, not a human. Secondly, iconicity is the only mode through which you can learn something about the object using nothing but the representamen, like we can learn something about humans from looking at a stick figure representing a human. However, this cannot be done by an icon alone (most signs are of a mixed type), because icons cannot refer to specific objects: the stick figure cannot refer to a specific human, or even a specific species — it might be an asari, or a turian. (Or a Presgr diplomat.) This has the simple but powerful corollary that if we can learn something about the object using only the representamen, then that sign must have an icon embedded in it. I'll go over how this works in a moment.

An index refers to its object by being spatio-temporally contiguous with it — that is, in the same place/vicinity as it in space and time — and thus it can "point" to or focus our attention on its specific object. This is a Secondness, involving a necessary relationship — spatio-temporal contiguity — between representamen and object. Thus, if I point my finger at my cat to draw your attention to her — notice that the cat in question has to be present at the time I point, and I would have to be within sight of her — then my pointing finger is an index of my cat. Likewise, smoke is an index of fire: it is spatio-temporally contiguous with the fire, being a necessary and intrinsic byproduct of it.

An index, then, focuses the attention on a particular object — something icons cannot do, as anything that shares qualities with an icon might be its object. Moreover, because an index is spatio-temporally contiguous with its object, it is a sign that is "related to its object through co-occurrence in actual experience [...] The power of indices derives from the fact that the sign-object relations are based on co-occurrences within one's own life experiences, and thus become intimately bound as experience."⁽⁵⁾ The meaning of indices, then, is dependent on experience — the particular experiential history of the sign recipient. In this way, "The Star-Spangled Banner" may to different people — or even the same person at different points in their life — index baseball games, Olympic victory, or imperialism.

Finally, a symbol only refers to its object by convention or habit — there is no causal or natural (indexical) connection, nor are symbols related to their objects by any kind of resemblance or similarity. This is the realm of what we commonly refer to as "language". A common noun is a good symbol: there is no natural, factual, or inherent relationship between the word "dog" and its object. We merely associate the two by learned force of habit. Symbols are legisigns — they are of a general type, Thirdness — and so their objects may only be classes of phenomena, not specific instances. Thus, the word "dog" can only refer to the general idea of dogs; to get a reference to a particular dog we must provide an indexical element such as a deictic: "this dog". Unlike indices, the meanings of symbols are largely independent of personal experience and are relatively fixed through social agreement. So, even if I have never personally seen a dog, I will have some idea of what the word "dog" refers to because I have heard dogs described and talked about in my society.

The sociocultural context of a symbol doesn't have to be as general as "the English-speaking world" or "Western society". Sub-groups can have their own definitions. For example, "tech" means something different in the general English vernacular ("the tech industry") versus the pickup artist community (for example, negging) versus in the realm of speedrunning (a new glitch that allows bigger skips and shorter times, or a different way of moving the character).

These categories are not mutually exclusive; in fact most signs partake of more than one. With this trichotomy the categories do not "correspond directly to distinct, natural types of signs [but rather] to aspects of signs," and most signs involve a little each of iconicity, indexicality, and symbolism. Indeed, these mixed signs are often the most interesting⁽⁴⁾. I'm making this point explicit both because it's interesting and because it is often forgotten, and some critiques of Peirce are based on the incorrect idea that most signs are of pure types. We will be discussing iconicity (and indexicality, to a degree) in much more detail later, as this is where the realm of metaphor lies.

Trichotomy III: How Signs are Interpreted to Represent their Objects

As can be deduced from the increasingly lengthy headers, the trichotomies have been steadily increasing in complexity. Trichotomy I dealt with one thing only, the representamen. Trichotomy II dealt with the relationship between two things: the representamen and the object. Note how this adheres to the hierarchy of primary categories of Firstness and Secondness. We now turn to discussing a type of Thirdness, the trichotomy that involves all three elements on the sign-relation. Trichotomy III deals with how a sign is interpreted as representing its object; the mediated relationship between the object and interpretant. These can be seen as degrees to which the representamen can specify its object, which will become clear once we go over these.

A rheme radically fails to specify its object; it is interpreted as a qualitative possibility, a set of parameters. It represents a possible object and cannot offer information about real, specific objects. A rhematic sign is semiotically weak, and the work of determining the object is done by something outside the sign. For example, a pointing finger is a rhematic index. However, if I am pointing at my cat, my actual pointing finger doesn't tell you anything about the object. It's just a finger. It's the spatio-temporal context that does the work of determining the object: when I point at my cat, the only way to figure out what my finger is indicating — to determine the meaning of the sign — is to examine the context and discover that near my finger and in the indicated direction is a cat, which we can then take to be the object. Likewise, most words are rhematic symbols. For /dog/, the representamen tells us exactly nothing about dogs (it doesn't sound or look like a dog, for example), and all the work of determining an object is done by the socioculture that links /dog/ to the idea of dogs. A rheme, then, simply indicates the possibility of an object, and partakes of Firstness.

A dicent, on the other hand, can be seen as something that both points to its object and tells us something about it. This is because a dicent somewhat determines its object; it is understood to represent a specific actual existent object. Moreover, and most importantly, a dicent is interpreted as actually being affected by its object, thus being a Secondness. It is in this sense that a dicent can "tell us something about its object": because it is causally affected by its object, we can interpret it by reading back from the effect to the cause (object) that affected it. We'll come back to dicents in more detail in a moment.

First, let's look at where this leaves us vis-à-vis the previous trichotomy. By definition, then, all icons must be rhemes, as an icon does not, cannot, specify any singular, particular object. An icon may only give parameters, and any object that meets these parameters can be a referent of the icon. For example, a drawing of a crescent shape may refer to the moon, which looks like a crescent when partially occluded. Or, it may refer instead to a pair of horns, which also shares the crescent shape. Similarly, a clipart drawing of a light bulb cannot refer to any specific light bulb, but can only provide some basic characteristics of objects with which to associate it.

Fig 2.2: A clipart light bulb. It cannot refer to any specific light bulb, but just to light bulbs in general. (source)

Even a painted portrait⁽⁶⁾ does not refer to a specific person, but to any person that matches the depicted appearance, or to the possibility of some person appearing this way.

Fig 2.3: A photo of Keanu Reeves and a painting of Louis-Maurice Boutet. The painting is a rhematic icon. (source)

Likewise, rhemes cannot have a truth-value: they are not propositions but parameters. For example, "_____ is black" is a rheme, where the blank is unfilled. This obviously can't be said to be true or false or have any truth-value at all. It's simply a quality, not attached to any particular thing.

An index can also be rhematic. For example, the ring of a doorbell can be a rheme. Let's assume that I have a doorbell that can only be rung once and has a specific tune, so that the person ringing the doorbell cannot change how it sounds or repeat it to indicate impatience or excitement. The sound of that doorbell, then, points to some token of the type /person at door/ — and that's all it does. I can learn no more from my doorbell. For all I know, in actual fact a squirrel ran up my doorframe and rang the bell, or a fairy appeared, rang my bell, and disappeared. But to me, the doorbell can only say "someone at your door rang me."

Indexical rhemes, like iconic rhemes, have no truth value: "that cat" is a rhematic index (made up of symbolic parts, to be sure), pointing out a specific cat — and that's all. There is no truth or falsity to it⁽⁷⁾. Many of the symbols we associate with language — common nouns, for instance — are rhematic. So, /cat/ does not have any truth value, nor does it tell us anything about its object — whatever we know about the object of /cat/ comes from sociocultural knowledge and/or personal experience, not from the representamen.

Unlike an icon, an index can be dicent: both pointing out an object and giving us some information about it. A weathervane is a classic example of a dicent index — its orientation is actually affected by the wind, so it is in a causal relationship with its object. Thus, when we see a weathervane, we read back from the effect (the direction it's pointing) to the cause (the direction of the wind) and learn something about the object, the wind.

Another classic dicent index is a footprint on a beach, and it also helps illustrate another important principle of dicents: they always contain an embedded rheme. The object of the footprint is the foot that made it. It is indexical, having been actually in contact in space and time with the foot at the time of production, and refers to the specific foot that made the print. But we can also learn something about the foot from examining the footprint. This, of course, depends on who is doing the interpreting. You or I might be able to tell how many toes the foot had, or whether it belonged to a child or an adult. An experienced tracker may be able to additionally infer the weight of the person who left it, their gait, or whether they had a parrot on one shoulder. Notice that we are learning something about the object — this must mean an icon (a rheme) is afoot⁽⁸⁾ here. And indeed, it is so: the footprint certainly shares qualities with the foot that made it; it is an icon of that foot. However, notice also that anything we learn from the iconicity inherent in the footprint is general, a set of parameters: a person — any person — with this many toes and this size shoe and this gait and this weight.

In our weathervane example the icon is a little harder to pick out, but it is there in the similarity between the direction the vane points and the direction of the wind. Another good dicent index is a bullet hole. It was made by a specific bullet, and indicates that a bullet has been here; its shape is iconic of the bullet itself. Yet another example is a fingerprint. Indeed, much of science is based on dicent indices, and this is our primary means of learning about the world: through associations that also inform.

Dicent indices, then, are particularly important because they are direct and convincing signs: they are interpreted as real, natural, or true. Such a connection has often been proposed of body language: our unconscious gestures, stances, and expressions are taken to be dicent indices of actual internal states — they are taken to be the effects of real causes within our bodies⁽⁹⁾. And yet notice that body language can be feigned — this is commonly done by actors, performers of any kind, and politicians. This type of communication feels so convincing and powerful to us because we read it as "true", so when an actor or politician's body language is of the form that conveys chagrin, we are swayed by it. Of course, we are also aware that body language can be manipulated, and so we might not take it at face value. As soon something can be feigned or mediated — as soon as something we generally see as the effect of a cause is used separately from said cause — we pass from Secondness to Thirdness. We are no longer dealing with an index, but a symbol. Thus, a smile can be both an index of true happiness — as when it is an unreflexive, natural response to a real feeling of happiness — or it can be a symbol we use to convey happiness when we might not actually feel anything of the sort, as when we smile for a school picture. An act of lying or pretending, then, is one of producing an icon of a dicent index.

We see, then, that a dicent index is always "true" — by definition, a dicent index shows us the real effect of an actual cause. Of course, dicent indices can be, and often are, misunderstood or misinterpreted, but by definition the effect that should be read as the "meaning" of a dicent index has a true causal root.

Symbols can also be dicent, and here we see again that any dicent necessarily contains a rheme. By putting together our previous rhemes — "____ is black" and "that cat" — we can construct a dicent symbol, a proposition that tells us something but only means by virtue of cultural association: "that cat is black". Note that a dicent symbol may be false: that particular cat may in fact be white.

So that's rhemes and dicents. The third part of this trichotomy is the argument. An argument completely specifies its object; it is "a sign of law" (CP 2.252). A syllogism is a type of argument. Arguments necessarily contain rhemes and dicents (parameters and propositions). All arguments are symbolic legisigns. We'll be focusing on icons and indices though, so if you want more about arguments, ask me! Otherwise, let's see where we are now that we've laid out all the trichotomies.

Combining the Trichotomies

Now that we've gone over the basic Peircian framework, we can return to the three basic categories of Firstness, Secondness, and Thirdness. While all semiotic processes involve Thirdness (because a sign-relation is triadic), each of the classifications we have discussed moves from relative Firstness to relative Thirdness. So, the first term in each of the trichotomies (qualisign, icon, rheme) pertains to Firstness, the realm of possibility and quality, of the monadic mode. Likewise, the second terms in the trichotomies (sinsign, index, dicent) partake of Secondness: existent objects, experience, actual relations between objects. And, of course, each third term (legisign, symbol, argument) is in the realm of Thirdness: mediated, abstract, and general. Finally, the trichotomies themselves, as mentioned before, move from Firstness to Thirdness, from the monadic quality of the representamen to the dyadic relation between sign and object to the mediated triad of sign-object-interpretant (Fig 2.4).

Fig 2.4: The sign-relation and the three trichotomies. The sign-relation is shown with Trichotomy I aligned with the representamen or sign, Trichotomy II aligned with the sign-object relation, and Trichotomy III aligns with the sign-interpretant relationship. Each trichotomy is ordered from relative Firstness to relative Thirdness. From Turino (1999: 226).

Peirce's semiotic framework rests upon the combination of the trichotomies, giving rise to ten basic types of signs such as the rhematic iconic qualisign (like a feeling of "red"), the dicent indexical sinsign (like a specific footprint), and the rhematic symbolic legisign (any common noun, like /tree/).

This, then, is how words mean. Words are a collection of qualisigns manifested in sinsigns that are instances of legisigns. Most words are symbols, but some, like onomatopoeia, are iconic. Words can also have indexical properties, like deictics: "this", "that", and all pronouns are deictics.

So we're just about ready to talk about metaphors! Spoiler: they're a type of icon (the type of meaning Saussure found boring, come on). However, first we have to take a detour and look a bit at the nitty gritty of how exactly the sign-relation works. That will be the next post!

___________________________

For now, let's review again! There were a lot of terms this time!

Qualisign or Tone

A sign that is an abstract characteristic of feature. "Voiced" or "redness" are qualisigns, as is the pitch of a musical note. Qualisigns are culturally determined.

Sinsign or Token

A specific instance of a sign; a sign that is an actual occurrence or fact. Writing out the word "dog" or saying it out loud are sinsigns.

Legisign or Type

A law that is a sign; an abstract association independent of any particular instantiation. All sinsigns [dog] are instances of the legisign /dog/. Legisigns are culturally determined.

Icon

A sign that denotes its object by sharing qualities with it. An icon cannot refer to a specific thing, only to anything that shares qualities with it.

Index

A sign that refers to its object by being spatio-temporally contiguous with it, or by being actually affected by it. Indices can refer to specific, single objects.

Symbol

A sign that refers to its object by convention or habit. Most words, like common nouns, are symbols.

Rheme

A sign that radically fails to specify its object. Rhemes give sets of parameters. Icons, indices, and symbols can all be rhematic. All icons are rhemes.

Dicent

A sign that somewhat specifies its object; it both points to an object and tells us something about it. Icons cannot be dicent, but indices and symbols can.

Argument

A sign that completely specifies its object; a sign of law. A syllogism is a type of argument. Only symbolic legisigns can be arguments. Arguments necessarily contain rhemes (parameters) and dicents (propositions).

Thanks for reading!

Notes

1. My numbering of the trichotomies is based on that of Turino (1999). (back to text)


2. Note again that general knowledge omits the third part of this triangle. (back to text)


3. Turino 1999. (back to text)


4. Stjernfelt, F. (2006). Two Iconicity Notions in Peirce's Diagrammatology Conceptual Structures: Inspiration and Application, 14th International Conference on Conceptual Structures, ICCS 2006, Aalborg, Denmark, July 16-21, 2006, Proceedings. Lecture Notes in Computer Science, Aalborg, Denmark. (back to text)


5. Turino 1999:227. (back to text)


6. We are not considering the case of a labelled portrait that specifically tells us that this is a portrait of a real, specific person. In this case, the label, which is a symbol, is added to the portrait, which is an icon. Even then, a labelled portrait tells us only that this a portrait of a person with a certain name and a certain appearance — if someone else were to share those traits, the portrait could be of them instead. (back to text)


7. Notice that this rheme does not posit the existence of the specific cat. An index cannot exist as a sign without its object in the first place, so by saying "that cat" we are not saying "that is a cat" or "that cat exists" but merely drawing our attention to a specific thing, the specific cat. Neither does the legisign /cat/ in itself posit the existence of such a thing as a cat, in the same way that /unicorn/ or /aliens/ do not posit the actual existence of such entities. (back to text)


8. I'm so sorry. (back to text)


9. • For general kinesics: Bateson, G. (1972). Steps to an Ecology of Mind. San Francisco, Chandler Publishing Company.
   
   
   • For application to intonation: Bolinger, D. (1983). "Intonation and Gesture." American Speech 58(2): 156-174.
   (back to text)

Comments

[(Anonymous)]

This is a fascinating column! But if you don't mind, perhaps you could use a cut tag? These are long posts to scroll past when I'm trying to check the rest of my reading list.

[justira]

So sorry about that! It was an accident. The cut should work now.

[hrj]

Reading with fascination – I love how approaching the same topic (e.g., how meaning is constructed) through different lenses can spark new ideas. (And you’re doing a great job of leading me through the process here.) A few random reactions … I haven’t had a chance to work through the third part yet in detail:

Fascinating that, in this framework, the legisign is understood as a subjective/emergent thing constructed by the observer, whereas under a formal logic framework I suspect it would fall in abstractness in between the abstract quality and the specific token. I.e., that the concept of “red rose” would be considered to be a subset of “redness” and that specific flowers would be considered a subset of the concept “red rose”. Or, from another angle, that one might view specific flowers as the most objective concrete part of the framework, with the first level of abstraction being the category “red rose”, and the next level of abstraction being “redness” and “rosiness”. I suspect some aspects of the legisign approach might be more accessible in the case of more complex and more subjective type-categories. I.e., without the distraction of the possibility of defining the type by apparently “necessary and sufficient” conditions (e.g., “absorbs light wavelengths in the following spectra”, “belongs botanically to the following range of species”). This is touched on in your example of the three rose tokens.

In the definition of “index”, defined as requiring a spatio-temporal contiguity, my mind immediately leapt to a feature used in ASL (and I am exceedingly ignorant about sign language in general except in how I’ve been introduced to it via linguistics, so I don’t know the technical term for this feature). Basically, in a statement concerning multiple “third person” individuals or objects that are not present in the signer’s vicinity, those persons/objects will be…well, let’s say “indexed” to specific locations in the sign-space, and signs performed in that space (or directed toward that space) are understood as relevant to the “indexed” entity. So—very much the same underlying concept as how “index” is used in this framework but abstracted away from the spatio-temporal contiguity requirement.

I’ve been having fun with the concepts of reference/definition/specification in the context of my fiction, where the characters are grappling with the logical and philosophical underpinnings of a type of formal sympathetic magic. How do you create a “science” by which abstract ritual actions/utterances can be focused to produce specific, targeted results? And what are the ways the process can go wrong?

(By the way, I hope you don't mind the extensive comments here. I know how much work goes into this sort of post, and I know that reader engagement is like oxygen for the process!)