Hand drawing a game strategy with white chalk on a blackboard.

An Alternative View of Offense/Defense by Michael Bogaty

An Alternative View of Offense/Defense

by Michael Bogaty

with significant contributions by Rahul Gosain

In recent years, the community has adopted a near-universal paradigm for evaluation of arguments: offense/defense. On any issue up for debate, whether a framework, contention-level argument, or theory shell, only offensive arguments hold weight. Defensive arguments can mitigate that offense, but are rarely enough to reject an argument. “Risk of offense” has become a common refrain on almost every layer.

Offense/defense has led to many of the problems that currently plague debate. While it’s true that debaters should offer offensive reasons to vote for them, the paradigmatic view that most defense is merely mitigatory and that a risk of offense can be sufficient is problematic.

Ultimately, we think that judges should evaluate offense/defense in a different way. To win an argument, debaters must answer and defeat responses that indict the validity of their link chain. This does not eliminate the distinction between mitigatory and terminal defense; it merely changes it to something more logical. Any argument that indicts a fundamental premise of an argument is terminal defense and, if won, reduces one’s credence in the argument to 0%. An argument that lessens the severity of an impact is mitigatory defense and plays out exactly as offense defense does now.

I: Three Problems with Offense/Defense

1 – Discouraging Intelligent Responses

Since under offense/defense, mitigation can’t win the round it often doesn’t justify the time investment. This is part of why, instead of making smart arguments that act as defense on the link story of a disad, debaters go for absurd impact turns. The minute and a half spent on the impact turn wouldn’t be more strategic than 30 seconds of smart analytics that point out the true flaw in the link story

Two other examples prove this point quite clearly. When someone argues that affirming leads to nuclear war or is premised on a racist ideology, a defender of the policy in the real world would respond with “no it’s not.” If that argument is won, it appears decisive. No one presumes the need to generate offense such as “your ideology is more racist” to defeat that objection.

If the intuitive conception is carried into debate, debaters are encouraged to think about the flaws in their opponent’s logic, which leads to more realistic arguments engagement. While there may be nothing wrong with debate esoterica, it seems preferable for the method of argumentation to have some grounding in what real world argumentation is like.

2 – Removing the Connection Between Argument Quality and Strategic Value

Disads with tenuous link chains become strategic if they reach extinction since they can be made to preclude all other impacts even in the face of strong defense. Esoteric disads and disads with hard-to-turn links and impacts are strategic independent of argument quality[1].

Similarly, theory shells such as specification (“spec”) theory are incredibly strategic under an offense/defense paradigm since it is almost impossible to come up with reasons why it is proactively worse to specify a single thing, say, the funding of the plan.

The strategic value of these arguments is bad for the activity since they trade off with more substantive arguments near the center of the topic literature. While theory debate is not necessarily bad, these theory shells probably are since it is difficult to have an in-depth debate on these issues since speccing provides, if anything, a marginal benefit with few harms to weigh against. The truest responses are defensive ones, which are unfortunately excluded. The latter point is key: arguments become strategic not because they are good and have few compelling responses, but solely because they are difficult to turn.

3 – Constantly Encouraging Debaters to Uplayer

When there is an apparent requirement to generate offense on a certain layer of the debate, it often becomes more strategic to uplayer than to engage arguments. Against a util framework that claims that actor-specificity comes first and governments must use util, debaters are discouraged from spending time on why governments do not need to use util since these arguments are merely “defensive.” Instead, the more strategic opponent is encouraged to argue that other meta-ethical/ontological/epistemological comes before actor-specificity, despite the tenuous link.

With theory shells that are difficult to generate offense against, debaters are encouraged to uplayer with metatheory since they are difficult to beat under offense/defense. This is compounded by the fact that a “risk of offense” on metatheory is often considered enough to win the debate.

This constant need to uplayer also seems bad for debate since it prevents any real clash on an issue. Few would disagree with the claim that clash leads to more interesting and educational debates[2]. This constant uplayering also discourages creativity. There is little incentive to come up with an interesting and unique case when one’s opponent can so easily uplayer with theory or kritiks that cannot be defeated with true defense and make it nearly impossible for the debate to ever return to the substance layer.

Next, we’ll show how the paradigm could be employed in practice to solve these concerns and demonstrate why it is preferable to the current offense/defense model. Hopefully, these examples explain the way in which the paradigm would be applied.

II: Case Studies

AI Disad: Debater 1 reads a silly disad claiming that a living wage leads to robots being developed to replace humans which in turn leads to extinction when AI goes rogue. The disad has several missing links that Debater 2 points out. Debater 1 responds by saying that there is a risk that the missing links are true, and since there are no turns, there is a risk that the plan causes extinction. Debater 1 wins so long as they win a Bostrom-type argument.

With the new paradigm, if a link were found to be missing, then the disad should be given no weight. Any of the following arguments would constitute terminal defense

  1. There is no brink evidence for how a small change in the number of robots will cause the war.
  2. They never warrant how the specific robots developed as a result of a living wage engage in warfare – those would be military robots instead.
  3. The AI that the impact card discusses as going to war are a different type of AI than would replace low-wage workers.

Giving the disad any weight in the face of one of these responses seems unintuitive[3]. The debater reading an argument that a policy will lead to extinction has the burden of demonstrating that the change caused by the policy will actually be large enough to trigger the impact. Accepting a lack of any brink for an impact like economic collapse as terminal defense would discourage debaters from reading silly arguments that are not grounded in the topic literature and would be laughed at outside of debate.

Funding Spec: Debater 1 reads a theory shell saying that the aff must specify a funding mechanism for the plan. Debater 2 responds with a counter-interpretation stating that he does not need to specify funding, since checking in cross-x would solve all abuse. Debater 1 responds with disadvantages to checking in cross-x. Debater 2 makes compelling responses to all of those disadvantages that everyone in the debate agrees are quite accurate. Since there is no disadvantage to having a three second text in the aff, the judge votes for debater 1 off of a risk of a disadvantage to checking in cross-x.

On our paradigm, Debater 1 would need to actively win an objection to CX checks, rather than the absence of disadvantages. Debater 2’s compelling responses are taken as reasons why there is no problem checking in CX instead – terminal defense. If at the end of the round, Debater 1 wins responses to every disadvantage to checking in cross-x, the shell would go away[4]

K No Links: Debater 1 reads an imperialism kritik with a poorly justified role of the ballot claiming that debate needs to discuss issues of imperialism. Debater 2 makes compelling no-link arguments (Say that the plan instead funds NGOs and lets them decide how to use the money rather than imposing ideals) and responds to the role of the ballot by noting that the Debater 1 doesn’t justify excluding other impacts or ethical frameworks. Debater 1 doesn’t respond to the substance of these indicts but argues there is a risk that the plan is still imperialist and the role of the ballot is correct since there are no offensive reasons to prefer a counter role of the ballot. The judge votes for debater 1 off of a risk of offense to the role of the ballot.

Instead, a debater would need to win proactive reasons why the aff links to the kritik, not merely the possibility of a link and the absence of a turn. Debaters would be required to defend links to the kritik and employ relevant literature and details of the plan to answer objections. This would enable debaters to win the link level (almost a lost art in modern LD) and concede to true arguments like “colonialist violence is bad.” This reduces the strategic utility of the kritik to a level commensurate to its applicability. As it stands, debaters have a similarly difficult time winning no link arguments reading a “bomb and militarily occupy Iran” aff as they would with an aff that funded grassroots movements. This is, in a word, insane.[5]

III: Responses to Potential Objections

1. Rejecting offense/defense leads to lots of blippy defense

First, it doesn’t seem that distinct from the status quo. Currently, plenty of rounds are decided on dropped blips anyway, and the status quo includes lots of terrible, blatantly ridiculous disads, kritiks, etc. Blippy defensive arguments will likely be more intuitive and help cut through the swath of bad arguments in the status quo.

The “blips” that would supposedly be read are also probably no warrant arguments and fallacies that opposing arguments commit. This would just encourage debaters to justify their argument in the face of holes being poked in it. A quality argument should withstand these responses, and debaters are free to note their relative underdevelopment as a reason to prefer their syllogism, which we imagine judges would be receptive to.

2. Every debate goes to presumption

This objection is simply false. While it’s possible that a slightly greater number of rounds would be decided on presumption, most debaters should still be able to win offensive reasons to vote for them. A debater unable to win an argument that can withstand their opponent’s responses probably doesn’t deserve to win in the first place – this is an unlikely scenario in most rounds. Further, it’s unclear why voting on presumption is any worse than voting for either debater on an argument that is thoroughly responded to. At that point, both routes seem pedagogically vacuous.

Also, if this is the result of rounds given what debaters are reading, there would be an incentive to move to fewer, well-developed, literature-based arguments that are easier to defend and have higher quality warrants. It’s unclear how that could be a bad thing.

3. Debaters no longer compare arguments

Just the opposite. Debaters will be forced to make higher quality arguments and compare the ones they’re winning, as well as compare arguments to defensive responses (which almost never happens now). If, in a DA scenario, one debater argues that warming is not anthropogenic, the other debater will need to compare the quality of their warrant and author qualifications instead of going for “risk of offense” arguments.The only arguments that will no longer be compared to each other are arguments that are proven to be fallacious or to have missing links, since these arguments will be given no weight instead of being assigned a risk of offense. The value of comparing arguments shown to be almost certainly false is quite low.

4. Skeptical arguments become far more common

While it is true that claiming a risk of offense against skeptical arguments will be more difficult, this objection is far from a problem with the model. Instead of making purely debate functional arguments about how “skep is just defense”, debaters will instead be forced to read philosophical literature about how their ethical theories can withstand skeptical objections, or arguments for why a small degree of credence in an ethical theory is sufficient to act. This captures the intuition of “risk of offense”, but allows for debates about skepticism to be more rigorous and literature based. These arguments are certainly more interesting than “skep is defense, vote on a risk of offense,” which, again, is only functional in debate.

Further, while not everyone likes skepticism, the fact that a class of arguments is allowed under a model for debate should not be a reason to reject that model. If skepticism is such a bad argument, it should be easily defeated. Under any model, skepticism only leads to a ballot if it is won. At worst, skepticism may be theoretically objectionable, which would eliminate this entirely.

Finally, debaters are still free to argue that framework debate is comparative, and so the better justified theory should win out rather than merely one that’s better justified. That seems like the norm in LD today, and doesn’t rely on offense/defense.

5. There is no non-arbitrary way to determine what is terminal and what is mitigatory

This is addressed earlier in the article. The distinction between the two is relatively simple: terminal defense points out a flaw in the logic of an argument while mitigatory defense accepts the argument as a logical possibility and attacks its probability or magnitude. Examples throughout the article illustrate the difference. Against the AI disad, for example, an argument that war is unlikely because governments could intervene against the robots is mitigatory since it concedes the disad is possible but merely lowers its probability. The other arguments mentioned above are all terminal since they point out flaws that would make the disad an impossibility.

It may be possible to rephrase analytics with stronger language, but that would also require strengthening the warrant. Recall that this paradigm deals with the implication of defensive arguments, rather than their structure.

[1] Indeed, they may be strategic because they’re low quality arguments. If they were better, they’d be written about in the literature and debaters could be expected to prepare responses. The need to have specific offense and lack of credence given to powerful defense means that debaters need to have those things cut in advance, which encourages debaters to flee to the margins of the topic.

[2] See Nebel, Jake. “Fear of Clash.” VBriefly. Victory Briefs, 1 Feb. 2013. Web.

[3] Certain defensive responses would still be mitigatory. For example, saying that AI going to war isn’t likely because governments could intervene merely lowers the probability of the disad and does not deny its possibility. The distinction should be clear. Terminal defense points out a flaw in the logic of an argument while mitigatory defense accepts the argument as a logical possibility and attacks its probability or magnitude.

[4] This doesn’t require a new understanding of competing interpretations. It’s still true that any offense to an interpretation is enough to vote. If a disadvantage to checking in cross-x (say its unclear what is said in CX), however small, is won, then the original interpretation remains preferable, so debater 1 would win the debate. However, expanding what counts as terminal defense means that an actual disadvantage to the counterinterpreation must be won. None of this excludes shells with a small amount of offense that is clearly won: that’s a question of competing interps vs. reasonability that this article doesn’t speak to.

[5]  It also has an odd strategic impact in that debaters aren’t incentivized to choose non-capitalist or non-colonialist advocacies, since those offer them little advantage in the resulting debate. Instead, they’re encouraged to read plans with big impact scenarios and quality solvency evidence (most often provided by big impacts) that will enable them to go for “case outweighs”-type arguments This may be a valid strategy, but it shouldn’t be the only option available.

  • Shmant the Warrior for Justice and Morality

    For a long time, my least favorite article ever written was the one that spurred Nebel T. After that, it was Branse’s article about truth testing, but now this one certainly takes the cake.
    I’ll go through each of your arguments against offense/defense and talk some why offense/defense is the best thing since sliced bread.
    Part 1
    1 – Intelligent responses
    To be honest, this just seems like a bunch of reasons debaters are bad at making strategic decisions. Link defense on disads is a great strategy since it opens up the possibility for weighing the case and it’s near impossible to win “risk of a disad” against a clearly won case flow. Mitigation is a perfectly good strategy in almost every instance as long as debaters can win some external piece of offense (including against arguments about racism and nuclear war). Also, it seems strange to say that we don’t think of “offense/defense” in the real world since that’s exactly how life works—absent an proactive reason to do something, we don’t do it and if there’s a proactive reason something is good/bad (aka offense), then we do’/don’t do that thing, respectively.
    2 – Arg Quality/Strategy
    Perhaps the problem with debate isn’t that extinction “can be made to preclude all other impacts”, but rather that people are just really really bad at putting defense on scenarios. I’m not sure why we should arbitrarily evaluate it as though there is ZERO risk of the disad once some link answer is put on it (which is what I’m pretty sure you’re saying we should do) because that’s just not how things work. Arguments aren’t made with full certainty, but rather there’s a possibility that each consequence will occur, which is why authors rarely use incredibly certain rhetoric.
    3 – Uplayering
    I won’t attempt to defend offense/defense (competing interps) on theory since reasonability seems like a good idea, which seems to resolve a lot of the uplayering issues, but I’ll talk about the framework stuff for a second. For example, on the “actor specificity” thing, I’m really not sure what the distinction between your paradigm and offense/defense would be. There’s not a “risk of a framework justification” in the same way that there is a risk of a disad because risks of disads are generated by some chance that a consequence will occur, which isn’t the case for frameworks. Additionally, people are still incentivized to engage the higher layers since they can generate offense against them (i.e. governments must use Kant/etc.) I also don’t know why that uplayering is bad—it seems as though if someone has a higher-level justification for some argument, then those would be applicable to the real world too (or those arguments are far less likely to be made).
    Part 2
    1 – AI Disad
    Arguments 2 and 3 are probably terminal (or really really close to terminal) defense under offense/defense too because they prove that the 1NC just hasn’t read a real link argument—it should be viewed as though they skipped from uniqueness to impact, which resolves your problem. The “no brink” issue seems to be resolved by this thing called “low risk”. Yeah there’s a risk of a disad, so if the aff wins no other offense, then vote neg, but otherwise it should be easy to weigh the aff against the tiny risk of a robots disad. It’s easy to just answer Bostrom (like seriously if you don’t have Bostrom answers you probably deserve to lose every round anyways). Also, this becomes a question of what should happen when they lose the case (since if the aff wins the case they win under both scenarios). To me, it seems a far more appealing option to vote on a tenuous risk of the AI disad than just presume.
    Actually, on the whole Bostrom thing, it seems like a lot of your arguments rely on a 1% risk argument being a large portion of offense in the 2NR/2AR, which is just how bad debate works. If the round becomes presumption v risk of offense, it should probably be decided on a risk. However, that should rarely happen because good debaters will actually make good enough arguments that they can win something outweighs a risk. Like congrats, offense/defense is terrible if people are bad at debate, but that’s not a reason we should dismiss the paradigm, it’s a reason people should stop giving bad 2NR’s.
    3 – K Links
    Pretty sure this is resolved by either epistemic modesty (@Bob) or just people properly understanding how the role of the ballot works. If they actually don’t provide a reason to actively exclude impacts that aren’t imperialism, then of course you get to weigh the aff against that. However, if the neg wins rejecting every instance of imperialism is key, then I have no problem with risk of a link being an option. It seems as though this model ASSUMES that other impacts should be weighed against the K (EM solves this) which is just an argument that can be justified in round.
    Part 3
    I don’t really think the first 2 objections are great anyway, so I won’t try to defend those.
    3 – Compare arguments
    This assumes that people don’t read defense under offense/defense, which I’ve already established isn’t true. Also, this assumes people will just only argue for risk of offense instead of actually debating, which is still strategically disadvantageous because it’s then harder to weigh it against other things (like the case). I think offense/defense is definitely best for comparing arguments—it forces people to defend any possible implication of the world they defend while still avoiding really trolly arguments by just giving them less weight.
    4 – Skep
    I mean, part of the reason skep is such a bad argument is that if it’s true, then it’s not like there’s any problem with doing the aff so we should just hedge our bets that it’s a good idea, which is how offense/defense works. Also, debates about skep are actually just ridiculous. Seriously. The race aff v skep example proves this—what educational model does it actually serve to argue about whether racism is bad or not??? Additionally, there’s even a good chunk of philosophical literature that says focus on skepticism is bad and we should just move past it. For example Wood and Roberts in “Intellectual Virtues: An Essay in Regulative Epistemology” argue that “some of epistemology’s most productive moments— in Aquinas, Kant, Plantinga— arose because philosophers were willing to set aside skeptical worries and look into what ordinary practitioners of science, religion, politics, and humanistic inquiry were willing to call knowledge.”
    5 – arbitrariness
    I’m pretty sure the distinction you’ve drawn is arbitrary. Just because you’ve drawn a line doesn’t mean that it avoids arbitrariness—why isn’t the bar for terminal defense higher or lower? Additionally, it seems as though “logical flaw” is pretty vague—even you examples would say that it’s “logically flawed” that war/etc will happen (in the case of the AI disad, I think). Just as there’s uncertainty in that argument, there’s uncertainty in other things you call “logical flaws”, which means it collapses back to offense/defense.

    • calm

      chill ok

      • Chill Pls

        this guy has absolutely no chill let me tell you a bunch of reasons why
        Part 1
        -Intellegent responses
        this guy has no chill
        Part 2
        -AI Disads
        You have no chill
        Part 3
        I don’t really think the first 2 objections are great anyway, so I won’t try to defend those. But you still have no chill…