 The Toolman Model of Argument, your map to writing essays as well as rhetorically evaluating the arguments of others. You've probably learned a lot of models for organizing and drafting an essay in the past, things like the Five Paragraph Essay are popular. However, as I've mentioned in past lectures, the Five Paragraph Essay is a really bad model to depend on, because it's entirely possible to fill all the requirements of a Five Paragraph Essay and not have an argument, much less an essay. Not something that qualifies for our definition of essay. Remember an essay is an ongoing process, it's an examination. It's weighing of different alternative interpretations of data and weighing the sort of conclusions that could potentially come from that. It's not just collecting a lot of information, putting it into a particular form and panning in a report or an information dump. And it doesn't just require investigating, researching and doing this thought process before you write. It is a record of that process that you can lead your reader through. And it's done within a particular context. The rhetorical situation includes not just the thing you're describing, but why it's important. Do you understand why this is important? And do you understand to whom it's important? Who is your audience? And your audience isn't just whoever ends up reading the paper, it's certainly not your instructor. The audience is only the people, the individuals to whom this information is relevant. And they have to be capable of understanding it, being influenced by it, but also doing something about it. And they can't just do whatever they want, they're under certain constraints. You have to understand what those constraints are. If we're talking about the way things ought to be at the global level or the national level, talking to individual people on the street may not help make that happen. But you have to know who your audience is in order to know whether something's important to them, how to communicate the importance to them, and then what they can do having heard your argument, having gone through that process of research and understanding with you. And any thesis statement you want to argue in an essay is a conclusion, and a conclusion has at least two premises as we learned from Aristotle. Aristotle was very helpful in helping us to come to the realization that data by itself, evidence by itself is never enough. A particular specific situation may constitute one premise, but one premise never by itself leads to a conclusion. We have to view it through the lens of another premise, some sort of general rule that helps us to understand specific data that we come into contact with. And to help us understand that Aristotle created the syllogism, the sort of abstract mathematical way to code the individual details of reality that we come into contact with. Unfortunately, reality doesn't always fit neatly into these sort of abstract molds. And even when it does, sometimes the realizations we come to are a bit too obvious. Or on the other hand, they're a bit too vague. That's why the philosopher and rhetorician Stephen Tullman, back in the 1950s, wrote a book called The Uses of Argument. He wanted to reintroduce Aristotle in the 20th century with an eye for practical application, because for the past thousand years people had taken Aristotle to ridiculously complex levels where they tried to, during the Renaissance and the Enlightenment era, tried to say that all reality can fit into this neat little equation or series of equations. And of course reality is just too difficult for that. Conversely, other people just wanted to focus on rhetoric for its persuasive value, where it doesn't matter how I lead an audience to a conclusion, it just matters that I do, that I persuade them. I can use fallacies, I can use overuse pathos or emotionalism. It doesn't matter. I don't even have to know what I'm talking about. I can be completely wrong about the thing I'm describing so long as I'm persuasive. Now that is the opposite problem. We have, on one hand, the search for truth that's too abstract, too mathematical. On the other hand, we have just sort of giving up the search for truth. No longer trying to make an argument of inquiry. Tullman wanted to give us a tool, a model, that would do both. And so he identifies six basic parts of an argument. And we're going to go through all of those parts, but these parts are going to be part of an argument or the understanding of an argument, whether or not they show up in the writing or in the speech. The first part's probably pretty obvious. That is the claim. The conclusion of your argument is going to be a claim. That is the thesis statement, the main thing you're trying to get your reader to accept. That's not going to be the only example of a claim. There may be claims that you have to establish within an argument before you can build on those claims to justify the primary claim. Tullman puts it like this. Let it be supposed that we make an assertion and commit ourselves thereby to the claim which any assertion necessarily involves. If this claim is challenged, we must be able to establish it. That is to make it good, to show that it was justifiable. In other words, we can't just say, here's this thing that's true, believe me. Most of us know this, we have to back that up. This claim is a judgment of some sort. So I'm going to use this icon of the gavel at any court judgment. A court judgment doesn't just happen by itself. There's always this evaluation of data that happens first. And Tullman uses the word data. We might also use words like evidence or sometimes he'll use the word grounds, the grounds for a claim, the support, the basis. And he says it this way, unless the assertion was made quite wildly and irresponsibly, we shall normally have some facts to which we can point to in its support. If the claim is challenged, it is up to us to appeal to these facts and present them as the foundation upon which our claim is based. And so remember what facts are. It's not just anything that's true in general. These are specific pieces of information that are independently verifiable. Let's say I point out the window and say that it's raining. You don't have to take my word for it. You can also look out the window to see whether or not it's raining. Now, a lot of times the facts are going to be pretty far removed. You're going to be depending on sources. So that your reader doesn't have to take your word for something. They can independently look up your sources, judge for themselves how accurate and well researched these sources are. But facts are out there independently verifiable. They're usually quantifiable. We can say how much of something, whether we're talking about global temperatures over the last 20 years or annual precipitation, the populations of certain species of fish in the Gulf of Mexico, whatever. But these are specific instances very often tallied up into a larger number. And when you write an essay you probably already know you're going to have to collect a lot of data. You're going to have to collect certain facts that are going to justify your main thesis. But you might run out of steam. You might sort of hit a wall after you gather lots of data. You might be wondering what is left for me to do. If all I'm doing is bringing these facts from other sources and making a claim about them, am I running the risk of just repeating the claim that other sources have made? How am I going to make my own argument here? And that's one of the things that the Toolman argument is very helpful for. For example, if I want to make the claim, you should take an umbrella if you go out today. I would back that up with data such as there's an 80% chance of rain. That 80% chance is or is not a fact, but it's part of the stasis of fact. It is something that you can independently verify. You don't have to take my word for it. So the data might be the barometric pressure and the chance of rain that that predicts. The claim would be because of this data you should take this action. But Toolman warns us even after we've produced our data we may find ourselves being asked further questions of another kind. We may now be required not to add more factual information to that which we have already provided, but rather to indicate the bearing on our conclusion of the data already produced. Colloquially the question may now be, what have you got to go on? How did you get there? In other words, we have to go from the data to the claim and sometimes that will seem obvious. And when we tell someone else the data, they will obviously jump to the same claim. But sometimes that's not going to be the case and we don't know whether that's going to work or not until we identify what it is that links the data to the claim. What general principle, what conventional way of thinking gets us from one to the other. So if I said there's an 80% chance of rain, that's my data and then therefore I make the claim you should take an umbrella, I'm skipping a step. There's a step I'm not mentioning. What connection does an umbrella have with rain? Now it's pretty obvious that umbrellas are designed to keep the rain off of you so I don't mention that in my statement, but I'm presuming that you and I both agree that umbrellas protect you from rain and since the umbrella protects you from the rain and there's a chance of rain today, you should take the umbrella. But when we're trying to make an argument that is not a given, that is one that our reader may or may not agree with, we don't want to take these sorts of steps, these sorts of connections for granted. Like Aristotle warned us, we just have one premise leading us to a conclusion. We always have at least two. So in this case the second premise would be that umbrellas protect you from the rain. And Aristotle calls this the major premise because it is one that we use not just in this specific situation on today when, you know, as the minor premise tells us there's an 80% chance of rain. The minor premise is very specific to a specific instance. The major premise is a general rule that we would apply in many different instances. And while it might seem obvious that, yes, umbrellas protect you from the rain, we could then say what other major premises might be applicable in this situation. Obviously umbrellas aren't the only thing that would protect us from the rain. A raincoat would also protect us from the rain. And if we're going into a situation that may include both rain and wind, such as here on the island, on campus, an umbrella might end up in Osso Bay. So we would want something that would keep the rain off of us even if it's coming in at an angle and won't blow away in the wind. So if someone disputed our claim, if someone said, I agree with your data, I agree that there is an 80% chance of rain today, but I'm not going to take an umbrella, we realized that the data by itself is not enough, that there is some unshared major premise. And so adding more data, adding more evidence that it will rain today won't change the person's conclusion or their claim. We have to identify what else does it work in that person's way of thinking. Another example I've used in the past is if you order a hamburger and you're given the option, would you like fries with that or would you like a salad? You get one side item, but you can't have both. You decide what you want, the fries or the salad. If I tell you that you should get the salad because it's healthier, I could give you lots of data about all the different vitamins and minerals that are in a salad, why they're healthy, why they're better for you, all the negative consequences of the fat, sugar and salt that are in the fries. I'm giving you data and I'm presuming that will lead you to a conclusion, but I'm skipping the step of mentioning that major premise that you should choose whatever contains the vitamins, minerals and fiber. But that's going to mean I might be overlooking an alternative warrant such as you should choose a side that tastes better. If my initial data about the vitamins and how they help isn't enough, I might throw on more and more and more data, but that's not going to change your mind if you don't share my warrant that you should get whichever side is healthiest. This is why I use this example and the umbrella example when we talked about the infamy. When we move from Aristotle's syllogism to practical application in a sentence, we don't always list every single premise that we need to get from the data, from the specific information to the conclusion, but it helps to do that when we're trying to figure out why we disagree or why we might disagree. And the term we use for this is warrant, and this is the same term that Stephen Tulum is going to use for this major premise. He says, supposing we encounter this fresh challenge, we must bring forward not further data, it doesn't matter if I just add more data about the healthiness of the salad or the likelihood of rain, that's not what's at issue. The propositions are of a rather different kind, rules, principles, inference licenses or what you will. Instead of additional items of information, our task is no longer to strengthen the ground on which our argument is constructed, but rather to show that taking these data as a starting point, the step to the original claim or conclusion is an appropriate or legitimate one. At this point, therefore, what are needed are the general hypothetical statements, which can act as bridges and authorize the sort of step which our particular argument commits us. So you can think of it as the chain that links the data to the claim. And, of course, a chain is composed of individual links and we'll have to make sure that each link is strong in that warrant. But identifying that warrant is the first step in a Tulum and Argument and it's a step that's all too frequently overlooked when people sit down to write an essay. People typically think, okay, here's my conclusion that I want to leave the reader to, so I'm just going to pile on all this information and assume the information speaks for itself. I'm assuming that the reader will make the same inferences, use the same general rules that I use when I look at this same data. And so Tulum starts off his diagram of his model of argument by having a D with an arrow from the D to the C, that is from the data to the claim, and then pointing to that arrow is the letter W for the warrant so that we understand that that warrant is holding up the arrow. Without the arrow, we just have data over here and a claim over here that aren't connected. In the past, I've described the warrant as an unstated assumption or unstated premise. That might be misleading because sometimes we do state it. Obviously in the Tulum and Argument, Stephen Tulum is arguing that we should make the warrant very explicit. So instead of saying it's the unstated premise, I'm going to call it the implicit premise. As he says here, the explicit appeal, that is the obvious, the stated appeal, the explicit appeal in this argument goes directly back from the claim to the data, relied on as a foundation. The warrant then is, in a sense, incidental and explanatory. In other words, I might not need to reveal it, but its task is to register explicitly the legitimacy of the step involved and to refer it back to the larger class of steps whose legitimacy is being presupposed. This is one of the reasons for distinguishing between data and warrants. Data are appealed to explicitly these facts. Warrants implicitly, most of the time. In addition, one may remark that warrants are general, a general rule, a general inference. You know, when you see this specific situation respond in this way, that would be a general warrant. It's a general rule certifying the soundness of all arguments of the appropriate type and have accordingly to be established in quite a different way from the facts we produce as data. So let's look at this example. This may be a familiar example. If you're driving along and you look at the horizon over a neighborhood, over a few houses, and you see black smoke billowing up, like lots of it. It's not coming from a chimney. All you can tell is it's coming from behind some houses. In that situation, the data is the smoke that's rising over the horizon. You say, I see smoke rising above a house. Now, I immediately make an inference about that. I make the inference that this is an emergency. There's a fire. The warrant for this would be, well, first of all, the obvious warrant would be where there's smoke, there's fire. But this goes beyond that. This goes beyond that to say that this is a dangerous fire. This is an out of control fire. And fire destroys houses. So not only does this smoke indicate that there's a fire, the fact that it's coming from somewhere that seems to be not a chimney and not where fire is supposed to come from leads me to think that this is a destructive fire. And I should probably call 911, call the fire department. And that warrant may be true some of the time under certain conditions. But someone else may ask, wait a second. I agree that where there's smoke, there's a fire. And I agree that fire destroys houses. But does that mean that this particular situation is an emergency? Is it possible that, yes, there's a fire and yes, fire is destructive, but this warrant doesn't necessarily lead us to the conclusion, to the claim that this is an emergency fire, a destructive fire. This question about whether the warrant actually applies to the data is what Toolman labels the rebuttal. Now, it's very important to clarify here that the way Toolman uses the word rebuttal, it's not just any counterargument. It is specifically a counterargument that looks at the warrant and says, yes, generally that warrant is true, and looks at the data and says, yes, I agree with the data, so they're not challenging the data and they're not challenging the warrant. They're just challenging whether or not that warrant applies to this specific data. Is it a valid warrant in this particular situation? Does it necessarily lead us from the data to the particular claim being given? So if we were to put that rebuttal into words, it might be that a contained and controlled fire, such as a barbecue pit or a leaf pile, could cause the same amount of smoke. If it's a lot of smoke, it might be burning leaves, burning trash, something like that, but that would not indicate that a house was burning or that some other property was being destroyed that wasn't intended to be. And this rebuttal is another sort of general question that the person making the argument might agree with sometimes. That might lead the person making the argument to say, well, yes, let's check that. Let's verify whether or not this claim necessarily comes from, is warranted by this data. As Toolman says, some warrants authorize this to accept a claim unequivocally, in other words, no matter what the situation, given the appropriate data. These warrants entitle us in suitable cases to qualify our conclusion with the adverb necessarily. This is very close to what Aristotle tells us about necessary conclusions where you don't need to make an argument. It's simply, take it for granted, one premise will always lead to one conclusion. But there are also arguments where you have to show the connection here. So here Toolman says this way, besides the necessary arguments, there are others that authorizes to make the step from data to conclusion either tentatively or else subject to certain conditions, certain exceptions, certain qualifications. In these cases, other modal qualifiers, such as the words like probably and presumably, are in place. It may not be sufficient, therefore, to simply specify our data, warrant, and claim. We may need to add some explicit reference to the degree of force which our data confer on our claim in virtue of our warrant. In a word, we may have to put in a qualifier. Now, you may have been told by writing teachers going back to elementary school that you should never hedge your claim. You should never hedge your thesis statement. Don't say something like, well, maybe this is the case because that sounds weak. A strong argument is where you say this is absolutely the case. However, you could probably think of a lot of strong arguments that were strongly inaccurate. You understood how the person came to that conclusion, but they seemed too fixed on that way of thinking. As Friedrich Nietzsche says, when the only thing you have is a hammer, everything starts to look like a nail. If you have this one particular warrant, then you might apply it in situations where it doesn't belong. So adding a qualifier, as Thulman describes it here, is a way of saying that this claim would only apply to this data under these conditions. And to know what conditions those are, we have to know what the rebuttal is. So if the rebuttal is a dispute that the warrant doesn't match the data, we have to understand what that rebuttal is in order to figure out what our qualifier should be. In this case, I would say that fire is destructive, fire would indicate an emergency, unless it was controlled, like in a barbecue pit or in a leaf pile. So I'm going to add that qualifier. Whether or not I actually make the rebuttal explicit, I do want to make the qualifier explicit that I can't see behind those houses, I just see smoke coming from where I don't think smoke should come from, but I realize that it might be a contained fire. There might be somebody tending it, but I want to find that out before I call 911. So the qualifier here would be unless the fire is contained and controlled, then there is a fire emergency. That establishes the conditions that if it is contained or controlled, it's not an emergency. It might still be a fire, but it's not an emergency. Tillman says in addition to the question whether or on what conditions a warrant is applicable in a particular case, we may be asked why in general this warrant should be accepted as having authority. In other words, somebody might not agree with the warrant at all. They might say that that warrant is not something that I should accept in any condition. In defending a claim that is we may produce our data, our warrant, and the relevant qualifications and conditions, and yet find that we have still not satisfied our challenger. For he may be dubious not only about this particular argument, but about the more general question whether the warrant is acceptable at all. Does not that warrant in its turn rest on something else? What he's getting at here is that a warrant is also a claim. If it's a warrant that our audience is going to share, then we don't have to back that claim up. We can just assume they'll already have this warrant. I just need to make sure that they understand that it applies in this situation. But there might be members of the audience or members of the readership that don't share that warrant at all. Someone might say in response, is that actually a general rule that fire just to destroy houses? You know, should I consider that as a rule by default? Or is that really just an exception to the general rule that fires a normal human technology that only rarely becomes destructive? Shouldn't we view smoke rising as a sign of normal activity, such as leaf burning or outdoor cooking, rather than as an emergency? This may sound similar to the rebuttal, but remember that in the rebuttal, the person making the rebuttal agrees that the warrant is generally true, but doesn't think it applies in this particular situation. In this case, someone is actually disputing whether the warrant is generally true. Is this a rule that we can usually assume? So I might have to back that warrant up. I might have to give more information that would lead us to accept this general rule. If you know people that are from a rural area where burning yard trash is not a big deal, you may find that they're surprised when they move to a more crowded urban area that there are ordinances against burning things in your backyard outside of an outdoor stove or something like that. So I might have to convince this person that yes, a fire is normally dangerous. You should normally accept that a fire is dangerous unless certain conditions are in place. So I have to maybe describe that there are city ordinances against outdoor burning because it is too likely that some ashes or some sparks from that fire might end up on something flammable and cause a neighborhood fire. I could produce more data that shows all the different instances in the past where a large blaze has burned out of control and burned up a house or a neighborhood that started as a fire of burning leaves or something like that. Now notice that I am giving more data, but this is not in the part of the argument that Toolman labels this data because if we think of the warrant as a claim that needs its own support, the backing would include specific instances. But in this case, I'm marshaling a lot of specific instances, a lot of specific facts in order to encourage the reader to make a generalization in the form of my warrant. So just because you read someone else's argument and you see the description of specific facts, that doesn't necessarily mean you want to label that as data with a capital D to fit in that formula. This might actually be facts that are relevant for the backing. So we see here the six elements of Toolman's argument. We're already familiar with at least two, the data, even if we refer to it as evidence or grounds or support or whatever, and the claim, this is your main conclusion. In this class we've talked previously about warrants, but notice that once we identify that warrant and make it explicit, we then know where to go to figure out would someone agree with the warrant in general but think it's out of place? If so, we can identify that rebuttal and from that rebuttal figure out what qualifiers we need to show that this is the appropriate data to apply that warrant to lead to the claim. But there are people who are going to dispute that warrant as a general rule, and for those people we need to add the backing, more support, maybe more evidence, more examples that would show that that should be a general rule. You should accept that as a general rule. So the data are the specific facts, evidence, grounds, which support a claim. The claim is that specific position on some doubtful issue, some controversial issue, or some issue that people don't already know about, that the arguer wants the audience or the reader to accept. The warrant is the general principle that connects the two. The backing is the support for why the reader should accept that warrant, and then if the reader does accept the warrant, you then need to understand situations, conditions in which the warrant would not apply so that you could add a qualifier that says, as long as this is the case, then this claim follows from this data. So let's take another example of how understanding this map helps the thinking process, not just in a single argument, but helps us figure out what to say next when we see an argument that has data, warrant, and a claim. I've used this example before, but in World War II, there was a situation where, because bombers were being shot down at a very high, unacceptably high rate, the crew on bombers in World War II had a 50% fatality rate. In other words, half of the people who went out on a bombing raid, or a series of bombing raids over a given period of time, would die in combat. And in order to prevent this, the Army Air Corps realized we need to add more armor to the planes, but keep in mind the bombers are already very heavy, so we can't cover the entire plane in armor. So where should we put the armor? What specific places should we put the armor? So they collated a lot of different schematics of planes, bombers that had returned from bombing runs, and they showed where the anti-aircraft bullet holes were. And they put them all together on one sort of diagram of a bomber, and it looked something like this. Now, no plane came back with this many bullet holes, but combined, these were the areas that always tended to show bullet holes. And that includes the tail section, the tips of the wings, and the center of the fuselage. And from this collection of data, the Army decided, okay, well, let's put the limited amount of armor that we can add to a plane. Let's put that on those areas that show that they're taking the most damage. Now, what they did here was they took a certain amount of data, and from that data, they jumped to a conclusion without necessarily examining the warrant, and examining potential alternative warrants. Luckily, though someone did, a recent European immigrant to the United States, named Abraham Wald, joined the Army's statistical research group. And he noticed that that warrant, that just because this is where the planes that return take damage doesn't necessarily mean that that's where we should put the armor. He saw that warrant, and he didn't necessarily dispute that warrant in general. The areas that receive the most damage should actually receive armor. However, he noticed that damage to these areas is not fatal to the aircraft. In other words, the aircraft with damage that's being recorded are the ones who make it back. What about all the other airplanes that didn't make it back? Damage to areas like the cockpit, where the pilots are, to the engines, which, you know, keep the plane moving, to the back part of the fuselage where the bombs are, the explosives. Damage there seems like it would be much more fatal, so it starts to make sense that we don't see damage there because the aircraft that took damage there would not have made it back. This is called survivorship bias. We're only gathering data from the ones that survived. So once he made that rebuttal to that warrant, that caused him to look at the exact same data but say that this warrant is incomplete. He changes that warrant to say that, yes, I agree that the planes are taking damage in those areas, but the fact that the bombers with damage in those areas were able to return indicates that these areas need the armor the least. Now he could back that up by saying aircraft with damage to the cockpit, to the engines, to the fuel tanks that are in the middle section of the wing and to the bomb payload would crash or explode in air and their damage would not have been recorded in that data set. So the previous warrant assumed that the data set was enough. Abraham Wald's warrant presumes that the data set is not enough. We might imagine a rebuttal to his warrant that would say, well, these other areas are important too. We can imagine a plane that loses the use of its outer rudder would have a hard time getting back and be vulnerable. But he would acknowledge that and say that, yes, if we could add armor everywhere we would, but as long as the qualifier is, as long as we are limited in the amount of armor we can put on the plane, let's put it in the areas not shown on the returning aircraft. Let's put it on the areas that are probably, where damage would be fatal to the aircraft. A past case study that we've looked at in this class and I've discussed in a past video is this article by Jonah Lehrer for Wired Magazine about the research of Robert Sapolsky who has genetically modified a virus that could get past the blood-brain barrier that's being tested in rats, but potentially at some point it could be used in humans that would stop the damage to the brain caused by glucocorticoids, these chemicals in the blood that are produced whenever we're under stress. But if you recall, that article is a very long one and it took Lehrer a long time to define terms like stress because people tend to jump to the assumption that they know what stress is. It's a feeling, and if you want to get rid of the feeling it's just a matter of calming down or taking your mind off something or getting some exercise, but he had to show that the subject that he's describing is not a feeling, it is something measurable in the blood. It is these glucocorticoids in the blood. And he makes his warrant explicit that if we have the ability to medically treat this damage that's done by stress, then we should do it. It is something we should research and develop. And the data being that there is something in the blood that coincides with stress that causes damage to the brain and the data that in mice, Robert Sapolsky has been able to minimize that damage through this treatment. And if we have this treatment available, we should use it, or at some point it will most likely be developed. That's his ultimate claim. And as he points out, many people, including people in the medical community did not take stress seriously. They didn't see stress as being a really a health issue or a problem. So they would dispute that warrant that stress isn't worth researching and developing. It's just a feeling. So he had to back up that warrant by citing data from Robert Sapolsky's research on baboons where he showed that the lower ranking males in a baboon hierarchy were less healthy, you know, more underfed, but also higher stress and more likely to have heart conditions. He compared that also with a study of the British Civil Service, another institutional hierarchy where people who were ranked lower were under chronic stress and four times as likely to die by the age of 60 as the people who were higher in the hierarchy and underwent only temporary stress. So he establishes his warrant that if we have the ability to fight the damage done by stress, then we should do it. He also realizes that some people might say, I agree that if there is damage done by stress, then it should be developed, but they might say that well, not all stress is bad. There are other means of reducing stress. You get stressed when you play a game, any kind of competition against someone else. You're highly stressed during it, but you lose that stress afterwards. And knowing that rebuttal is out there, he adds a lot of qualifiers. It's not just a matter of, you know, this will probably be available someday, you know, probably, mostly, or something like that. He gives many different qualifiers to show that this new treatment is not just for making people feel better. It's not going to alleviate the need for other things like relaxation therapies and, you know, it's not, you know, doing exercise is still a good way to fight stress. And he strongly recommends that we continue to reduce the social conditions that cause chronic stress in some people. But despite all those qualifiers, they don't alleviate the need for this treatment for people for whom we can't help, the people who can't just get rid of stress, the people who endure more than temporary stress, who endure chronic stress. There are, under certain qualifications, there are still going to be people who need this treatment. So he makes all those qualifications very explicit before telling us that, you know, this is why we need to have this treatment available eventually. And we probably, and hopefully you will remember from the case study, the way that simplifications of this argument skipped all of those qualifications in the backing, and just interpreted it, interpreted all the research as this scientist was studying how to make people feel better by just giving them a shot. But that completely missed what was actually going on in the argument and caused a lot of confusion. So Larry's article is a good example of a toolman argument that has all of these components made explicit. This is discouraged just because you read an article and you can't find examples of the qualifications or the rebuttal or the backing. A lot of arguments don't have that. They may not even make the warrant explicit. So once you know what the model is, if you come across an argument that doesn't have all of these elements, that in itself is very important. That's something you want to take note of, and if you're writing a rhetorical analysis, you want to say what is missing and what that could lead to, the article or, as Abraham Wald did when he analyzed the warrants of the aircraft engineers, point out that the warrant is missing or it's misapplied. So just because you can't sort of go through an article and label all of those elements, that doesn't mean that they're not relevant or that they shouldn't be there or that they're not necessary. So another case study is this article by this psychologist Stanley Corrin that was like John O'Leary's article simplified and repeated by many other articles. But his article in the magazine Psychology Today said that the data says, don't hug the dog. And the best version of his thesis or his enthamene there is that new data shows that hugging your dog raises its stress levels and its anxiety levels. And the data that he's referring to is a collection of 250 pictures like these two that come from his article where a human is photographed hugging a dog. Now he got these photographs from Google images and Flickr and that sort of thing. And this is data. But what he points out are things like the dog's ears are down or they have what he calls half moon eyes where we can see the whites of their eyes because their eyes are turned. Head is turned to avoid eye contact. Submissive eye closure which seems to be the opposite of the half moon eyes where they're open too wide. The dog is licking their lips. Head turns to avoid eye contact, ears down. So these descriptions are the data. But his warrant is this. The signs of stress and anxiety in dogs are well established and are easily observable at least by trained individuals. Obviously at the high end of stress we have dogs who bear their teeth but there are subtler indicators. Anxiety is when the dog turns his head away from whatever is bothering or worrying him. Sometimes also closing his eyes, at least partially. Alternatively dogs will often show what is commonly called a half moon eye or whale eye which is where you can see the white portion of the eyes in the corner or the rim of the eye. One common visible sign of stress or anxiety is when the dog's ears are lowered or slicked against the side of the head. Lip licking or licking a person's face can also be signs of anxiety as can yawning or raising one paw. These signs and other similar ones should be easy to detect in stress dogs. All that I needed to then do was to conduct the research was a source of photographic material showing people hugging their dogs. Now notice this is a long version of his warrant but he has the warrant before he has the specific data. To consolidate his warrant we can say that these cues that we see in the photographs the closed eyes, the wide open eyes, raised ears, lowered ears, lip licking and licking human faces and yawning and raising a paw. Those facts are signs of stress. It would be inaccurate to say that the dog's the data shows dogs showing signs of stress because sign of stress is an interpretation of what's happening. We don't want to overlook that. We don't want to mix the interpretation with the data. And you'll notice that his backing for this warrant that these cues are interpreted as stress. The backing is pretty thin. He says quote, the signs of stress and anxiety and dogs are well established and are easily observable at least by the trained individual. He gives a little bit more when he says that dogs are cursorial. In other words they're animals that like to run. But as any evolutionary biologist who's studied human evolution can tell you humans also evolved to run and we hug each other as a sign of affection. There's no rebuttal and there's no qualifier to the rebuttal to those who would agree that, yes those cues are correlated with stress in dogs but they might not apply to those specific instances, to that specific data, to those pictures. And because he omits that rebuttal he doesn't add the qualifier that might actually be pretty necessary. And this precise rebuttal is given in a subsequent article by Rachel Feldman at The Washington Post who says that okay the cues that are correlated with stress in dogs, she doesn't dispute that there is that correlation. But remember the difference in the causation status between correlation and causation. In her article she says she goes over the way science works with an experimental group and a control group and you submit this description of your research of the experimental and control group to a peer-reviewed publication and then other experts in the field look over it to make sure you've taken all the necessary steps. This was not done in this case. Psychology Today is a magazine, not a peer-reviewed journal. And in particular this jump in logic from the specific data gathered to the interpretation that dogs feel stressed when hugged or held is not necessarily warranted. Even though that warrant that these signs can be associated with stress, they might not necessarily in those specific situations be associated with stress. Feldman says a peer reviewer might argue that Corrin had no real information about the environment in which the photos were taken and the emotional state of the dog beforehand. They'd want to know there was some kind of control group which in this case would mean a way for scientists to see what each dog was like when the variable of the hug was removed. In other words, if this little girl on the top left or this guy just to the right, if they weren't hugging those dogs, what would the dogs be doing? In particular, the dog on the top left is described as looking away from the little girl, but notice it's looking at the person taking the picture. And that's the problem with the photograph as evidence. You don't know what happened just before the picture was taken. What's happening outside the frame that could cause the same reactions or the same actions that the dogs are showing. Not to mention that if you own a dog or you're familiar with dogs, you know that when dogs lick your face, that doesn't necessarily mean that they're stressed. It could just mean that they like you. If you play tug with a dog and it looks to the side while it's having fun, you're going to see that half moon eye as Corrin describes it. Dogs with heavy ears frequently have them lowered. And of course, yawning, there's a lot of reasons for yawning other than feeling stressed. So in other words, these cues might be correlated with stress, but that doesn't necessarily mean that stress caused them. And Corrin, the author of the first article, is presuming that there is a causal relationship. His warrant is not necessarily a justified interpretation. It doesn't necessarily apply to those photographs. We don't know if it applies or not, because we don't know what happened before the photograph was taken or what's happening outside the frame of the photograph. So this is why the Toolman model of argument is very helpful with doing a rhetorical analysis of someone else's argument to see if there are important steps that are left out. Of course, like I said, not all arguments will have every one of these elements explicitly described. And if they don't, that might be a good indication of what's left for you to do in your analysis. But for the same reason, it's also a great model for using to generate your own essays. If you're not sure what to do after you gather lots of data, think about what interpretation you come to from the data and then ask yourself how you got there. What's your warrant? And are there alternative warrants that might just as well explain the data and lead to a different claim? What would that be? Figure out what the rebuttal is. Not just any counter argument, but specifically someone who agreed with the warrant in general but thinks it doesn't apply to the specific situation. And then what would you have to say, how would you have to add something to your claim in order to say as only under these conditions, only with this qualification would my claim be valid. But I think my claim is valid because I think these conditions are met. Like with the example of seeing the fire coming from over the horizon or over the rooftops of houses, you might have to make a claim before you know all the data. So that's where the qualifier really helps you say, this seems to be the case and if it is the case then this claim is valid but it might not be the case. Don't let somebody tell you that that's hedging, that you should never hedge, never make a weak argument. The weak arguments are usually the more articulate arguments, the more reliable, the more cautious arguments. Scientific arguments are going to be full of qualifiers like this because in science you don't want to overstate what your data show. You don't want to assume that a dog that's licking your face is stressed because there are, you know, other explanations for that same data. And then there are going to be people who don't share your warrant so you're going to have to make your own sort of miniature argument to support that general warrant before you can then use that warrant to interpret the data. And I know it might seem just as difficult and abstract and everything as Aristotle's syllogism and the Anthememe but once you learn how to look for these different elements in another article or figure out what you need to do in writing your own you're going to make both rhetorical analyses and writing your own essays much, much easier, much more intuitive. And you're not going to run out of things to say after you describe the data you've collected.