I'm a sandwich descriptivist rather than a prescriptivist.
I know what a sandwich is, and any definition that say that a hot dog is a sandwich or that a sub/hero/hoagie is not a sandwich is a definition that I do not accept.
Similarly, a bagel with cream cheese (that is, both halves of the bagel sandwiching cream cheese) is not a sandwich (despite the use of the word "sandwiching" in the definition), but a bagel with turkey inside is a sandwich.
I'm comfortable without a definition, though I look forward to hearing the reports of the intrepid theoreticians that can yield definitions that satisfy these criteria, much as I laud the efforts of the adjective-ordering folk in descriptive grammar.
My rule of thumb is that if I went into a restaurant and asked the waiter "I want a sandwich, but I don't care what kind -- surprise me!" and they brought me an object, how angry would I be at the waiter who brought me that object.
> any definition that say that a hot dog is a sandwich or that a sub/hero/hoagie is not a sandwich is a definition that I do not accept.
Isn't that still prescriptivism? Descriptivism here would mean noting what people claim constitutes sandwichhood, and maybe noting what people's behavior reveals about what they actually do or don't treat as a sandwich. But it would be prescriptivist to assert a new definition of sandwichhood, even if you argue that it's somehow more motivated than others.
By descriptively I mean that I do not believe that we should model our usage after rules which were derived from usage, but instead should examine our usage and attempt to model it.
A prescriptivist says that a pronoun should match the number, gender, and human-ness of its antecedent, and therefore “they” is not acceptable when referring to a single person. A casual descriptivist says that they understand what it means when other people use “they” in a singular fashion, and find that it is a useful construct. A formal descriptivist would try to describe how the usage and clarity is practically evolving over time. The terms “pronoun” and “antecedent” are terms themselves modeled from actual usage, rather than some organizing committee that decided how English should work.
Similarly for sandwich definition. The descriptivist viewpoint says that no strict definition is necessary; that people have a casual understanding and should model their usage on that. A prescriptivist says “here is a definition of ‘sandwich’ that models a significant number of things that many people agree are sandwiches, so we expand that definition and use that to label things as sandwich or not sandwich, regardless of the common understanding”.
The key point here is that I do not feel a strict definition is necessary in order to reasonably discuss sandwich labeling.
Practically speaking, I have found that my heuristic above generally yields agreement from others (in particular, I have not yet found anyone who would be happy with a hot dog under those circumstances), with occasional differences of opinion that would be interesting for a professional sandwichologist to examine as a question of sandwichial drift over time.
Prescriptivism versus descriptivism is ultimately about what is more important: is the set of rules primal, or are the labeling of examples primal? Both people would agree that an utterance like "dragon they q" is not grammatical. But when faced with an utterance such as "me and John went to the park," the descriptivist would say that it occurs frequently enough that it ought to be considered grammatical, and that we should understand what the properties of "me" are to allow it to be grammatical, whereas the prescriptivist would object that "me" is an object pronoun and therefore cannot appear in a compound subject phrase so that people who utter such a sentence are not uttering grammatical sentences.
Or put differently: if reality and models differ, a descriptivist says that the models are wrong, while a prescriptivist says that reality is wrong.
I am a scientist and clearly am earning millions from Big Sandwich to define the applicability of sandwichness to hot dogs. Please do not subpoena my email.
> My rule of thumb is that if I went into a restaurant and asked the waiter "I want a sandwich, but I don't care what kind -- surprise me!" and they brought me an object, how angry would I be at the waiter who brought me that object.
Despite the fact that you're clearly a mental patient who thinks that a bagel with some meat inside can be a sandwich, you have my upvote for that line:-)
Seriously, though, that's actually a pretty good criterion. Reminds me of Justice Potter Stewart's comment about knowing it when he sees it: if I ask for X, someone gives me Y & I'm angry, then Y is not in the class of X to a degree directly correlated with my anger at the result. I dig it!
> Similarly, a bagel with cream cheese (that is, both halves of the bagel sandwiching cream cheese) is not a sandwich (despite the use of the word "sandwiching" in the definition), but a bagel with turkey inside is a sandwich.
What if instead of cream cheese it is a slice of lettuce, some cheese, and some mayo, for a veggie sandwich?
Remove the lettuce, now what?
Does lettuce, green colored water, define it as a sandwich? Or is it the presence of a solid type of cheese versus spreadable? Is brie + lettuce in a bagel a sandwich while brie alone is not?
Remove the cheese, just leave the mayo. Now what? Add back in the lettuce (say the eater is lactose intolerant), does it become a sandwich again?
> What if instead of cream cheese it is a slice of lettuce, some cheese, and some mayo, for a veggie sandwich?
I’ll go ahead and stop you here; no, what you have is not a sandwich. By which I guess I mean that it is not an unqualified sandwich in the same sense that an open-faced sandwich is not a sandwich. If I directly ordered either nobody would think me odd, or if the waiter asked me “is a veggie sandwich okay” I wouldn’t think him crazy. Oddly enough, a grilled cheese would be fine as an unqualified sandwich.
I’ll turn this around and ask you to do the thought experiment. If you were the waiter at a lunch counter and someone ordered a sandwich and said they didn’t care what kind, would you give them a veggie sandwich, playing the odds? Would you as a customer feel okay getting that response? Or another similar thought experiment of your own devising where you would feel comfortable with that usage?
Plenty of places sell vegetarian sandwiches. 50 cents of thin sliced turkey does not a sandwich make.
> Would you as a customer feel okay getting that response?
I've had some great veggie sandwiches, the one I described above wasn't a great example, it was purposefully minimalist to prove a point.
Shove some sliced grilled eggplant on there and it'd all be good.
And how in the world are you OK with a grilled cheese being a sandwich but if someone pokes a hole in the bread and uses cream cheese it now is no longer a sandwich?
Is a grilled cheese sandwich with a circular hole cut out of the center a sandwich? What if it isn't grilled? I'm trying to understand where is the line is here!
For what it's worth, my entire point is that there is no line. A line is completely unnecessary. Just like a native English speaker doesn't need to be told what the rule is for adjective ordering -- "the big red dog" sounds right, "the red big dog" sounds wrong.
I've had perfectly fine veggie sandwiches, when I've ordered vegetarian sandwiches, but that wasn't my question -- would you, personally, consider that to be a reasonable response if you went into a random restaurant, or if you were asked to prepare a sandwich for a customer who said they didn't care what kind of sandwich?
I'm not asking rhetorically -- I'm genuinely curious. People do differ on this; many people I talk to disagree on the bagel question, and consider both variants a sandwich or consider neither one a sandwich. My theory here is that I grew up eating bagels at home and not at restaurants, so would eat my bagels with cream cheese open faced, and not sandwiched between the two halves, so I never really created the association with being a sandwich; while people who consider it a sandwich grew up in cities where you got a bagel from a store or deli, where it is pretty much universally served with the two halves stacked, and thus is more sandwich-like during developmental years.
Would you feel comfortable, without knowing the customer, giving them such a minimalist veggie sandwich and still expect that they would be satisfied that you've obeyed the broad parameters?
I believe the term you're looking for is "particularist", by analogy with moral particularism:
"Moral particularism is the view that there are no moral principles and that moral judgement can be found only as one decides particular cases, either real or imagined." https://en.wikipedia.org/wiki/Moral_particularism
I think I have seen a proof that any algorithm which decides whether or not something is a sandwich in general can be used to decide whether or not a turing machine halts on no inputs.
If I went to a restaurant and ordered soup (surprise me!) and they gave me a bowl of cereal, I would likely have my goons come back and torch the place.
I mean, if they gave me gazpacho or a cold melon soup I also would not be thrilled, but I would not accuse them of violating the conditions that I had laid out.
I suggest training a tensorflow model with foodstuffs labelled by a large group of, say, 100,000 human volunteers, carefully selected to be demographically representative.
Given that you're clearly aware that different people have vastly different definitions of "sandwich," how could you justify being angry at all if all you ordered was any unspecified sandwich?
Yes, different definitions, but most people in usage don’t start from a definition and proceed to usage; they start from usage. In this case, I am explicitly rejecting the notion that there is a definition that matches usage (both mine and others).
My experience, anecdotally, is that given the rule of thumb above and similar thought experiments, there is not a lot of variation in the practical usage of the word sandwich.
Opposition mainly comes from people who attempt to apply an arbitrary rule and are not willing to revise that rule to match their own usage for the sake of the argument. In other words, overthinking. Or a pathological case of Occam’s razor.
Or put it in a cuboid New England roll or a roll more typically used for sandwiches as would be needed to wrap larger hot dogs and sausages. The problem is really the mass market minimalist hot dog roll.
Using my rule of thumb above: if I asked the waiter for a grilled cheese, but I didn't care what kind, I expect the waiter would be somewhat perplexed by this, since "grilled cheese" is a sufficient descriptor that there's basically one sandwich that satisfies the description.
If I asked for a melt but didn't care what kind, then I feel like the possibilities expand, and would include a regular grilled cheese.
> Using my rule of thumb above: if I asked the waiter for a grilled cheese, but I didn't care what kind, I expect the waiter would be somewhat perplexed by this, since "grilled cheese" is a sufficient descriptor that there's basically one sandwich that satisfies the description.
Even if you take grilled cheese narrowly enough that it must have nothing but cheese between the bread, there are a fairly wide array of potential breads and potential cheeses and combinations of cheeses that there is definitely more than one sandwich that fits the description, even before considering that, in use, “grilled cheese” allows for additional ingredients.
As someone who regularly eats "grilled cheese with stuff on it" (bell pepper slices are awesome), the distinction hinges on whether the cheese holds the sandwich together. If the halves can be easily separated, it's not grilled cheese.
The Portuguese standard here is a grilled cheese and ham toastie. Just cheese is less common but popular-ish, heathens (e.g. me) might ask for just ham.
True, there's a wide space of breads and cheeses, but I can walk into a restaurant and order a grilled cheese and with no follow-up questions be served the restaurant's version a grilled cheese. The same could not be said of ordering a "sandwich"; the waiter would certainly demand clarification before accepting the order.
The conversation could go like this, "I'll have a grilled cheese" -- "What kind of cheese?" -- "surprise me" -- "What kind of bread?" -- "surprise me" -- "Do you want tomato or bacon or any other meats?". At this point I'd have to stop and figure out which way I wanted to go. But that's reasonable because these are non-default options; if I asked for a surprise sandwich and the waiter asked me "is open-faced okay" or "is ice cream sandwich okay" or "is a wrap okay" then I'd totally accept that as a clarification, but without that clarification none of those would be acceptable.
This gets us classifications all the way from "no starch at all" (i.e. soup) to "non-starch surrounded on all sides by the starch" (i.e. calzone). But I posit the existence of an eighth category... that is to say the inverse of the calzone, which would be "starch surrounded on all sides by a non-starch"? An anti-calzone, if you will. Seems rather rare, but not impossible (e.g. turkey with stuffing).
Once, while working in a pizza place I theorized the creation of what I like to call the Dyson Pizza. Sauce, cheese, and toppings completely encircling a sphere of bread.
It's suddenly obvious that not only could this be baked in space, it absolutely will be. This will be a thing when space tourism comes online. I only hope they still call it a Dyson pizza.
But what if it is a donut that is both filled and glazed? Is the glaze just on top or all the way round?
What does a beef Wellington count as? It has multiple nested layers... what if you made a turducken and wrapped each bird in a flatbread of some kind, before stuffing it in a larger bird?
Some of the post's salad examples (spaghetti, poutine) could be interpreted as starch surrounded by non-starch. Thus I believe the rule would still consider turkey with stuffing to be a salad.
It’s not a salad if it’s a single cohesive unit. If you chopped it up, it could pass as a “turkey salad” I guess, in the sense of a potato salad, but that would still be a stretch.
Yes, Spaghetti is misclassified as "Salad". It is a solid starch entity where if you sampled any part of the cube, you would get some starch. A loaf of bread, a baked potato, etc are all anti-calzones.
It seems we might slip from group theory to category theory. At which point we are surely describing the structure of meta-sandwiches (in that they sandwich various definitions of sandwichness between each other)
Sandwiches are like fascism. In Ur-Fascism, Umberto Eco talks about the difficulty of defining the "know it when you see it" category of fascism. We start with a set of concrete examples and we seek to define the category based on the qualities shared by those examples. One example might have qualities A and B, a second B and C and a third, C and D. The first thing shares no qualities with the last, but they all belong to the same category by a kind of transitivity. Sandwich is an agglomeration of qualities that tend to cluster together.
> Still, just as we cannot give a final, essential definition of ‘game’, so we cannot find “what is common to all these activities and what makes them into language or parts of language” (PI 65).
> There is no reason to look, as we have done traditionally—and dogmatically—for one, essential core in which the meaning of a word is located and which is, therefore, common to all uses of that word. We should, instead, travel with the word’s uses through “a complicated network of similarities overlapping and criss-crossing” (PI 66).
> > ... We should, instead, travel with the word’s uses through “a complicated network of similarities overlapping and criss-crossing”
This is untenable. It's part of the identity of certain words that they change meaning, but that's the exception. In all other cases you can find a center of mass, so to speak, although describing that with fever words than the word you try to describe is mostly impossible. I need to be able to reject when somebody talks bullshit, like Wittgenstein does there, professionally.
It's absolutely untenable for the practical purposes you're after. Unfortunately, that doesn't disqualify it from being true. The more we examine language, the more it seems to be the case that there's really no guaranteed 'centers of mass', that they can shift between corpora, that they are different between times, geographies, social registers, and even between individual people. Some semantic units may be more volatile than others, but they're still all defined relative to each other within the head of a given speaker, and non-rigorously at that. And on top of that, they've got pretty complicated, non-rigorous connections to whatever the driving hardware underneath looks like.
It'd be nice to have some formal process for deciding what's bullshit and what's not. But, it doesn't seem to actually. exist in any objective sense. The words that make the most sense to you are your own. And everyone else always kinda seems like they're full of bullshit.
It's the same with "knowledge". Classical knowledge is understood as justified true belief. But you can come up with countless counter examples where someone is justified in believing something that is indeed true but you would not attest that person any knowledge of the fact. [1]
In Czech Republic, we have "sausage in a roll", which is the same thing as hotdog, except they make a hole in the roll (from one side only) and stick the sausage and sauce in it. So you can nicely hold it in your hand like ice cream cone.
Recently, some companies here have started making classic hotdogs. But I never understood why - "sausage in a roll" is so much superior topology (for eating on the run). On the contrary, I don't understand why Americans won't adopt "sausage in a roll" instead.
Corndogs are a similar concept and are common as fairground food, but with a cornbread rather than wheat bread layer.
For a traditional hot dog, though, the key thing is the toppings. Some American styles, like the Coney Island dog, can have as much topping as there is hot dog.
In Denmark, this is called a "French hot dog". There isn't agreement whether it's actually inspired by something similar from France, as it appeared first 30+ years ago.
Interestingly both Germany and Poland have something similar but under different names. The German version was invented by a government food research center in an attempt to replicate Western foods: https://en.wikipedia.org/wiki/Ketwurst -- so you are right in that the top DDR food scientists also agree about the ideal hotdog shape!
For quite a while, fast food in Denmark was all about a mobile "sausage stand" (the permissions to which were interestingly given preferably to those who could not have a full time job).
In certain parts of America, we have. Klobasniky (most people just call 'em kolaches) are incredibly popular in Texas, perhaps even more popular than hot dogs.
The fun part, is that in theory, the klobasniky came from the Czech via the kolache, and then came back to Czech republic from Texas filled with meat. =)
I remember getting American-style buns and the rolls that you're referring to. I prefer the American-style buns because when you take a bite from the roll version, the excess sauce drips out the other side, whereas you can maneuver an American-style bun such that it doesn't drip.
If it's done well, it won't drip out. See the post about Ketwurst above, that's exactly what I mean (and I didn't know it was - it seems, might be a myth - invented in GDR).
You carve out a hole slightly larger than the sausage. You hold the roll at approximately a 45 degree angle, while spinning it. Thus the hole remains in place, while the bread moves around it. As you are spinning it, you put the toppings in. Then you shove the sausage in (after acquiring enthusiastic consent).
Sauce-like toppings (mustard/ketchup/etc) are essentially put in one end before the sausage, and as the sausage is pushed in, it spreads the sauces throughout the whole sausage.
Some others (fried onions, etc) get put in with a long spoon before the sausage.
One I enjoyed was a discussion on a gaming podcast about whether fireballs in street fighter were a form of magic or not, eventually leading to the question "is Guile a wizard?"
Welcome to the wonderful world of Ontology, "the philosophical study of being, [...], existence, reality, [...] and the basic categories of being and their relations". [0]
A narrower related field is Taxonomy, the practice and science of classification.
I like the "cube rule" as a very coarse grained way of classifying food! Classifications of things can become pretty crazy, in particular when they try to classify very general and dissimilar things (example, the schema.org vocabulary [1]).
Now someone should turn the cube rule to an OWL vocabulary so machines too, like we humans, can understand the difference between a hot dog and a salad [2].
I would argue that sandwiches follow a prototype semantic representation in American English. That is, there are a handful of "ideal" sandwiches which are the first thing you think of when it comes to a sandwich. Some items are conventionally not sandwiches. These items might have many similarities with things which are sandwiches, but because of their cultural or culinary significance, are in their own special category. A person may judge an object based on their relative sandwichness which is really a distance metric in which items closer to one of the platonic ideals of a sandwich than they are to a not-a-sandwich are considered to be a sandwich. This way of mentally representing categories of objects, while natural, makes forming precise, rule-based definitions of sandwich classification inherently difficult and inconsistent. Any algorithm for classifying a sandwich must either contain rules which are not universally applied in all situations, or which classifies (or fails to classify) foods as sandwiches which are not conventionally considered sandwiches.
As a non-sandwich example of the same phenomenon, consider whether or not an item is considered to be furniture. A few items are universally considered to be furniture: beds, couches, desks, wardrobes. These items together form a semantic boundary, but that boundary is not necessarily uniform. Having legs certainly matters, but not all furniture has legs. Being not easily movable matters, but chairs are easily moved and if couches are furniture, so are chairs. Likewise, items which are not movable at all, such as shelves built into the wall may not be furniture depending on who you ask. Appliances have electronics, and appliances could be considered a special category of furniture, or simply not be considered furniture. Some people would classify lamps as furniture, but not televisions.
The issue is that semantic boundaries are not symmetrical. Having features which differ from our platonic ideals and our prototypical example of the category are not weighted equally. Some features matter more, and all of those features together form a gestalt of sandwichness. An attempt to form hard and fast rules about which features are necessary to be a sandwich are well-meaning, but ultimately incorrect as they fail to appreciate the psychology of the sandwich, what it means to hold the concept of a sandwich in your head.
The way I see it, anything between two other pieces of another thing (separate pieces or conjoined, doesn't matter) constitutes a sandwich. If we accept the origin story (urban legend?) of its origin as invented by the Earl of Sandwich so that he and his friends can snack without getting their hands dirty during card games, then the sandwich can't be defined as a specific food item, but by the act of putting one thing between two others.
Furniture is anything used to "furnish" a space. If the shelf was attached to the wall after the wall is built, it's furniture. If it was built as part of the wall, it isn't.
There are a few missing variations that aren't symmetric to one of the 6 existing categories. I wonder if this implies that there are other categories possible that haven't been made yet.
The one with the most possibility is two starch faces, as in (2), but the starch faces share an edge.
I immediately thought of this with the pumpkin pie slice.
Also, the site indicates that a whole double-crust pie is a calzone, while a slice of double-crust pie is a taco. Having food categories change as they are served and consumed seems fundamentally unstable and can lead only to chaos. The single slice could easily pass through three or four stages during consumption alone! I think a food's category must be fixed at the end of preparation, for the sake of all sanity.
No, there is no acceptable state between burrito and salad. A split taco is an initial demonstration of innate knowledge and some motor control, best done before the staff retreats.
Yeah, also it may seem obvious that a bagel sandwich sliced in the traditional would fall into the {top,bottom} Sandwich category, but they provide no formal descriptions of smoothing. I'm sure I could construct some wildly-non-homotopic constructs that would defy the "obvious" smoothing.
(Is an uncut bagel a Toast or an "air" Sushi?)
If I were arguing for this rule, though, I'd say that the unlisted combinations should collapse to the listed ones based on curvature. {side,bottom} collapses to either Toast or Taco, and {side,back,bottom} collapses to any of Toast, Quiche, or Taco.
A Scotch-Egg is a very curious exception to the cube rule. It's a kind of calzone, where edges have been limited, or a kind of 3D toast fepending on your topplogical analysis...
I think the real mistake is assuming that the answer to “is this a sandwich” is either “yes”, or “no”. Something can be kinda a sandwich, or not really a sandwich, or slightly a sandwich, or definitely a sandwich, or not at all a sandwich, or anywhere in between.
Furthermore, there doesn’t have to be a consistent set of rules defining what is and what is not a sandwich.
This is because the categories are meant to capture some information about an object, but we expect them to capture all of the information about an object.
My friend has a deceptively simple answer to this question:
If you put schmear/spread/sauce on the bread, it's a sandwich.
If you put schmear/spread/sauce on the "meat", it's not a sandwich.
The distinction is that in the latter, the bread is just a vessel to get "meat" into your mouth. In the former, the bread is part of the whole package.
The hotdog is classified as taco because it has three sides of the cube but I would argue that a hotdog consists only of two perpendicular sides of the cube because the angle between the buns is only about 90 degrees.
I subscribe to the belief that the key feature of a sandwich is it's horizontal alignment and stacking. Therefore a hot dog isn't a sandwich because it's split in the middle, not stacked horizontally.
Actually subs are the reason that the horizontal/vertical rule needs to exist-otherwise we could just use two separate pieces of bread as the rule. A sub is stacked vertically whereas a hot dog is nestled horizontally.
Ok, i tried to take it serious, but i'm dying here of laughter.
I am still convinced this is meant in a quite serious way, bbut. Gneehehehhahah.... FLAG IT!
I understand that many of the examples are just absurdism in action, but I have trouble understanding the resistance to labeling a hot dog as a sandwich.
But on the other hand, my wife likes to troll me by calling buttered toast a "butter sandwich", so maybe some rigor in our definitions would do my life some good.
I shall not today attempt further to define the kinds of material I understand to be embraced within that shorthand description of "sandwich", and perhaps I could never succeed in intelligibly doing so. But I know it when I see it.
If anyone is wondering why toast (1) is listed as an example of a sandwich (2)... I recently learned from a kind HN commenter that toast sandwich is a sandwich where the filling is a buttered piece of toast.
The definition that I have always stuck with is that a sandwich has leavened bread on two sides of whatever is in the sandwich. Therefore, hot dog is a sandwich, a taco is not, etc.
Interesting. It all comes down to topology - of the wrapper (or cover) and the fillings. BTW, what happens if the wrapper itself has filling, for example falafel wrap?
Lasagna does not have pasta (its starch component) on the bottom or the top. A wrap (such as a falafel wrap, as it is pictured on the site) is a taco, not sushi.
Right, but that's not the point. I am referring the sushi standard defined in the article based on the cube illustration. In reality there is sushi that conforms to nearly all of those designations. Furthermore, tacos are filled, not wrapped.
Crepes are not trivial to classify: obviously a plain flat crepe with some topping is a toast, a "walkaround" (usually sweet) crepe is a quiche and sushi/wrap crepes exist (e.g. galette-saucisse[0]) but what about the "complète" where the crèpe is folded back on the filling but not entirely? Functionally it's closer to a calzone or a toast (it's not eaten from the "opening") but technically it looks more like a quiche.
Most of the sandwich debate is caused by ignoring adjectives. In the case of a submarine sandwich, it's a sandwich except we apply the adjective submarine to say we are modifying the bread situation. An open faced sandwich is just like a sandwich except we remove the top piece of bread. Which is why a hotdog is not a sandwich. We could have an adjective to say we are modifying the bread and another to say we are modifying the filling, but then there is nothing left of the sandwich portion, which is why we have a new word for it.
Whether you call it sandwich or not, the mathematical description of a hot dog is well within the bounds of the mathematical description of a sandwich.
Fiddling about with adjectives is about as useful as saying that squares are not rectangles or rhombuses because otherwise they would be called rhomboid rectangles or rectangular rhombuses.
The hot dog is a special case of a sausage submarine sandwich.
My only objection is the somewhat capricious way the rule is reapplied based on serving. The author classifies pumpkin pie as an instance of toast, not quiche, because the photo they found was of a single slice, while cheesecake, deep-dish pizza, and key lime pie are all labeled quiches because the photos are of an entire pie. An Uncrustable is even labeled a calzone only when “unbitten” because it apparently becomes a quiche the moment you bite into it!
While this may be topologically consistent, it’s not useful as a taxonomy to have foods dynamically change categorization as you serve or eat them.
The only thing I'd change is to rename the sushi category to something broader like wrap. Sushi is a really specific foodstuff to me to be the name of that category.
A bunch of croutons on the bottom of a plate look like toast, but you can't lift them all by picking up just one. Unless you're cooking your mashed potatoes in a way I've never seen, you won't be able to pick them all up by lifting one part of the mashed potatoes.
That brings up an even more interesting/disturbing question does a burrito turn into a salad after the first couple bites because one (hopefully only one) end is open now.
According to the cube rule it would turn into a soup/salad w/ bread bowl. Tacos have 3 open sides a partially eaten burrito has one open side. Burrito -> taco requires a massive structural failure.
Super-sets/classes - A calzone is still a sandwich since it implements the sandwich class (distinct bread on two sides of a food; which could be an additional slice of bread. E.G. Folding a slice of bread over half of a different slice of bread can be a sandwich. However making an S-curve does not cause a distinct middle layer (though if the edges are sliced off suddenly it is). ).
I have a passing interest in linguistics[3] and so let me tell you, after intensive meditation about the topic I found that sandwich could well be cognate to German "Wecke", a bread bun, supposedly related to 'wedge'. I can't pin point "sand-" yet.
I think that would be crucial to determine the precise meaning. I'm very much against either extreme, pre- or description, but I find that expanding the sample space back to include historic meaning may yield results able to satisfy both sides.
Maybe I can still help out. We are actually talking about the club sandwich. Club sandwich could be, for all I know, which is very little about the club sandwiches history, cognate to Ger. "Klappstulle" (a 2-sandwich [to use the OP's typology] made by folding a slice over) at least the club part, and both could go all the way back to the Proto Indo European root whence glue (to stick together) compare Ger. "Kleber" (glue), "Kleie" (wheat-paste?), and perhaps Old English and Gothic "hlaif" (bread) then most likely in reference to dough, Ger. "klumpen" - 'clump', 'glob'. A club of people sticks together. The swinging and clubbing club could perhaps compare to 'key' (from Lt. clavis, from a sense nail, peg, hook), as stick, sticky and clavis, claudo show similar relations. "Klappe" means 'hatch', 'shutter', 'mouth', we say something "klappt nicht" (doesn't function), "(um)klappen" ... it's baffling that there's no good translation: The lid of box is a "Klappe", and closing it shut is "zuklappen"; A "klappe" that doesn't shut tight, flaps and makes noise, that noise is called "klappern" and from that all kinds of 'to make noise' "klappern"; Best fit would be 'clamp'. Now it depends at which time, if there was a relation at all, our dinners aligned or diverged. But, since you can't take a Hot Dog to work and eat it later--already cold--it cannot be a club sandwich, at least if folding is done primarily to stack and wrap them without making a mess of it. Well, I made a mess of it, I didn't check references on most my claims, but I'll have to wrap it up.
As a working hypothesis I suppose "sand" relates to Ger. "Kanten", the thick end of a loaf, related to "Kante" (edge), but its Greek root has no secure origin so a link to the celtic word for rim, tire giving "waybread" or "wagonbread" (cp. Ger. "Weg" - way; "Wegebrot") is not too promising. It's just, I have that image of the thick end bread as a happy hobo's ration very present from cartoons. But a better candidate would relate to butter perhaps. Or how about "Shank", cog. Ger. "Schinken" (ham), "Schenkel" (thighs), as something two sided, angled. Anyone reminded of John Connery? That's ridiculous. How about "to send for" - Ger. "gemachte Brote". I mean, there was a time and place where ... wasn't common and bread had been eaten raw or at least separately (... "Belag" - topping, coating).
[2] from Latin canthus (“tire, edge of a wheel”), from Ancient Greek κανθός (kanthós, “corner of the eye, tire, etc.”). The Greek is not certain, maybe from Celtic or if that seems implausible not Indo European.
[3] While I don't like anyone prescribing as if they were they majesty, I still dislike the descriptive approach that labels anything as correct that randomly comes up in the mash up of dialects. I'm especially weary of misunderstandings becoming fashionable.
the word "send" doesn't seem to have changed much; Meant travel, journey. Sandwich was a landing site for ships, roman beach head and all. So sandwich was kind of ... a postoffice? traveler's office? Just a marketplace and harbour? I mean, compare Stralsunt and tell me that's named for being built on sunt, err I mean sand.
On the other hand, the look and feel of hardtack (which cannot be Tolkiens waybread if that's brown on the outside and creamy on the inside) and zwieback pretty much reminds of sand. And they might sell it to ships leaving port. That doesn't quite figure with how we know a sandwhich today, but sure enough, there are other roots that can derive "sand-".
At any rate, placenames lend much better and hence more often to food names. And placenames in turn have often been related to milling and possibly mill-derivatives; to biblical proportions--cf. Bethlehem--which, again, may mean that bread words overlap with more figurative meanings; compare "sound" (healthy), from PIE reconstruction *swent-~sunt- (vigor); Then compare vigor and -wich ... Townsville might sound redundant, but there it is.
reading these comments is weird. i interpret this page as a kind of satire of a modern pretentious analytical techie orientation that hubristically "solves" things by creating some alluringly simple model and ignoring all the ways it doesn't work. it adopts the mores and tone of a modern tech talk, and presents its idiotic results with a straight face. so its super funny. a little less funny when you have to explain it to people. people who are taking this at face value- really?
The problem is one of significance to machine learning, I believe. 6 DoF for the classification of food within the domain of dough products is ... seemingly too little.
It isn’t good to see large communities of humans burning mental calories to come to the same set of conclusions (I’m on team sandwich because $REASON | I’m on team taco because $REASON). It is a mental D.o.s. directed towards humans that tend towards herding.
Similarly, a bagel with cream cheese (that is, both halves of the bagel sandwiching cream cheese) is not a sandwich (despite the use of the word "sandwiching" in the definition), but a bagel with turkey inside is a sandwich.
I'm comfortable without a definition, though I look forward to hearing the reports of the intrepid theoreticians that can yield definitions that satisfy these criteria, much as I laud the efforts of the adjective-ordering folk in descriptive grammar.
My rule of thumb is that if I went into a restaurant and asked the waiter "I want a sandwich, but I don't care what kind -- surprise me!" and they brought me an object, how angry would I be at the waiter who brought me that object.