Reconciling Tech Over Truth

Introduction

“Tech over truth” is a phrase frequently touted on judge philosophies, at this point becoming a philosophy in and of itself. This philosophy has a common consensus, though the true nature of what a “technical” judge is remains unquestioned.

Colloquially, a technical judge is one who does not let their personal belief in arguments influence their willingness to vote for said arguments. For instance, a technical judge who personally believes that global warming is truly bad would still find themselves compelled to vote for Warming Good if won on the flow. This contrasts with judges whose personal disbelief in the benefits of global warming would cause them to disregard the argument, even if won in a debate. Other commonly touted examples include spark/wipeout, “process” counterplans, and the kritik (or untopical affirmatives).

However, an understanding of what exactly it means to be a technical judge leaves much to be desired. What does it mean for something to be “won” in a debate? Is it the stance of the tech > truth crowd that claims do not need warrants? How does a technical judge adhere to their principles while evaluating various edge-case scenarios that a non-technical judge would merely be able to exclude, citing personal belief?

I hope to reconcile some of these issues and present what I believe is a model for understanding technical and non-technical evaluation.

A Sentential Logic System for Debate

To understand how a technical judge should evaluate arguments, we’ll start by creating logical representations of all primitives relevant to a debate round. The most basic currency debaters leverage are claims. These are ubiquitous in debate, and debaters frequently claim that various premises are true. We’ll represent claims by using letters, for instance:

$$\begin{array}{rl}c& :\text{The plan solves war with China.}\\ q& :\text{The plan is unpopular with voters.}\end{array}$$

We can also display the negation of claims

$$\begin{array}{rl}c& :\text{The plan solves war with China.}\\ \neg c& :\text{The plan does not solve war with China.}\end{array}$$

Claims can have truth values as well. Here, I’m not referring to “real world truth” but rather whether a claim has been won as being true in debate or not. We can represent the truth value of a claim using a simple function:

$$T(x)=\{\begin{array}{ll}1& \text{if }x\text{ is true}\\ 0& \text{if false}\end{array}$$

Justifying Claims

Merely having claims is of little use to us. The degree to which a claim has been warranted seems relevant. Therefore, we need a way to represent this; we’ll use another function:

$$J(x):\text{The degree of justification for }x$$

For instance, if a debater has claimed

$$p:\text{The plan is popular with voters}$$

and then cites reputed polls conducted recently with high sample sizes, we can assign some arbitrary non-zero $J$-value to $p$

$$J(p)=5$$

Introducing justification enables us to break ties in the instance of contradictions. We can update our function $T(x)$ from earlier to map to the sign of the justification differential $J(x)-J(\neg x)$ as follows:

$$T(x)=\{\begin{array}{ll}1& \text{if }J(x)-J(\neg x)>0\\ 0& \text{if }J(x)-J(\neg x)\le 0\end{array}$$

If the degree of warranting for the negation of a claim is equal to or higher than the degree of warranting for the claim itself, then we can assign it a $T$-value of 0. For example, if a debater were to claim

$$m:\text{The plan gets modeled internationally.}$$

but provided minimal warranting, we could assign

$$J(m)=3$$

Now, if the opposing side were to argue

$$\neg m:\text{The plan does not get modeled}$$

and were to provide peer-reviewed evidence, backed by experts in the field, citing empirics, we could assign some higher $J$-value to $\neg m$ such that

$$J(\neg m)>J(m)$$

therefore concluding that $T(m)=0$.

Edge-case: Unwarranted claims

What to do with two equally unwarranted claims? Say a debater blurts out

$$c:\text{The plan prevents war with China}$$

and provides no warrant for the claim. Now their opponent also blurts out

$$\neg c:\text{The plan does not prevent war with China}$$

and provides equally little warranting. In this case $J(\neg c)=J(c)$. As the judge, do you conclude that the plan prevents war with China or that it doesn’t?

Using the above function for $T(x)$, you would conclude that since $J(\neg c)=J(c),T(c)=0$; therefore, the plan does not prevent war with China.

Real-world truth

For the technical judge, the above is a sufficient toolkit, and $T(x)$ can be determined solely through the arguments that a debater has made in a round. However, for judges that ascribe to other philosophies: truth > tech, tech + truth, “Bayesian,” an element of a claim’s connectedness to real-world truth---or personal belief---also matters.

We can represent this personal belief in an argument using another function

$$B(x):\text{One’s personal belief in the truth-value of an argument}$$

This value can implicate $T(x)$ in many ways. Some judges are extremists; unless they personally believe in an argument, they refuse to vote for it (e.g. parents). For them

$$T(x)=\{\begin{array}{ll}1& \text{if }B(x)>0\\ 0& \text{if }B(x)\le 0\end{array}$$

Others include some baseline check for real-world connectedness in order to evaluate arguments, for instance as a Bayesian lens

$$T(x)=\{\begin{array}{ll}1& \text{if }J(x)-J(\neg x)>-\log \left(\frac{B(x)}{B(\neg x)}\right)\\ 0& \text{otherwise}\end{array}$$

Here, personal belief is used as a threshold that an argument must overcome in order to potentially have $T(1)$. The possibilities are endless, though this is also the reason why such a model of evaluation is undesirable. This post is not intended to be a defense of the tech > truth paradigm, but it seems self-evident that debaters want the debate to be about the arguments they’ve made and not their judge’s personal belief, which including $B(x)$ anywhere in your assessment risks. Therefore, to truly remain technical, $B(x)$ should never be a factor in your equation for determining $T(x)$.

Reconciling Justification

One final bit of reconciliation is necessary, how do we determine $J(x)$? It seems that to include personal belief in evaluation runs contrary to the tenets of technical evaluation. However, this seems inevitable. You need to know what a PhD is in order to quantify by what amount $J(x)$ should increase if the author a debater is citing has a PhD. You also need to know about the real world in order to assess the quality of analytic arguments or evidence, etc. Otherwise, you’d have no way to distinguish the arguments a debater has made from gibberish.

This is compatible with a technical philosophy. For a judge to be technical, they must be willing to vote solely on the arguments made in a debate; i.e. they must not include personal belief in an argument, $B(x)$, in their evaluation of it.

A non-technical judge imposes some baseline, influence, or amplify $J(x)$ based on $B(x)$. The distinction lies in whether one’s knowledge of the world is used to compare warrants (is this evidence more qualified than the other) or claims (does the plan prevent war with China).

Distinguishing $J$ and $B$

A concern that arises is that $J$ and $B$ seem quite similar. I’ve conceded earlier that external knowledge is relevant for assessing $J$, but also seemingly that it shouldn’t be permitted via $B$. So what is the difference?

The distinction lies in that $B$ is a measure of pure personal belief, whereas $J$ is a measure of the justifications presented in a round. This means that $\forall x$ such that $x$ is a claim made / contested in a round, $J(x)$ at the beginning of the round is not equal to $J(x)$ at the end of the round. However, because $B$ just reflects your own personal belief, the arguments made in a round---unless they’ve been so compelling as to shape your real belief---never affect $B$.

Another question that arises is, what does it mean to permit personal belief in evaluating justification, but not in influencing the outcome? To answer this, we need a model of the technical judge.

A Conclusive Model of The Technical Judge

From all of the above, we can construct a model of a technical judge. We know that a technical judge must not let their external knowledge influence their willingness to vote for certain arguments. That establishes that the following are true

$$\forall x,y\left(\right[J(x)=J(y)]\to [T(x)=T(y)\left]\right)$$ $$\forall x,y\left(\right[J(x)\ne J(y)]\to [T(x)\ne T(y)\left]\right)$$ $$\neg \exists x,y\left(\right[B(x)\ne B(y)]\to [T(x)\ne T(y)\left]\right)$$

It should never be the case that a mere difference in $B$ implies a difference in $T$ for any two claims made in a debate round, and it must be the case that $J$ directly tracks with $T$, and therefore that differences in $J$ must lead to differences in $T$. We’ve already established the distinction between $B$ and $J$ earlier, so we’re left with the conclusion that, for a technical judge:

• Belief in claims relevant to a debate must be based solely on the arguments made in a round
• Knowledge of the external world should be relevant solely to assess the justification for arguments made, but must never be used to filter arguments that do not align with one’s personal belief
• Arguments made in a round should always take precedence over belief