In the Fall of 1982, at Amherst College, I (Milt Williams) had the privilege of taking a Philosophy class, Logic, with Prof. Joseph Epstein. I found the class fascinating from the very first day, as we learned about logical implication—the notion that P logically implies Q if, and only if, Q must be true if P is true. This logical relationship is similar to the relationship expressed by a conditional statement—if P, then Q—for if the conditional statement is true, and if P is true, then Q must be true (by modus ponens).

I never anticipated that the dynamics of logical implication and conditional statements would structure my thought process as a lawyer. I started my legal career at the New York County District Attorney's Office, where I had the opportunity to conduct numerous cross-examinations—in most cases, having the bare minimum of material, and in some cases none, to understand what a witness's testimony would likely be. I was forced to imagine the strongest story (from the defense perspective) that a witness (in many instances, the defendant) could manufacture. I then had to make a plan to deconstruct it. The goal of this exercise was to identify the propositions the witness was most likely to assert; the causal, factual, or logical connections the witness was most likely to assume; and the conclusions the witness was most likely to draw or be drawn into. Armed with this understanding of what the witness was compelled to say (given his or her prior statements and interests in the case), I would identify the points I wanted to elicit that would help me persuade the jury of the prosecution's case. Moreover, by thinking through the assumptions and logical implications of a witness's likely testimony, I learned that some of them were positively helpful, and thus I learned that cross-examination can have a constructive as well as a destructive purpose.

I did not initially realize that my preparation for cross-examination applied many of the lessons I first learned in Professor Epstein's Logic course. Because the witness could be forced to admit certain propositions having the logical form "if P, then Q," or because I could rely on the jury recognizing the obvious truth of "if P, then Q" or an obvious inference from P to Q, I could use the rules of logical inference to conceptualize my strategy and, in court, to elicit propositions that I could pull together and present to the jury in closing as a tight logical inference. See John Henry Wigmore, Principles of Judicial Proof 270 (Little, Brown & Co., 2d ed. 1931) (discussing the role of presumed "general truths" or "general principles," often conceived in conditional form, in our reasoning at trial). For example, a witness might testify that he saw P, and have to admit (on cross) that if he had seen P, then he would also have seen Q. By testifying he had seen P, he thus becomes committed to Q (by modus ponens, if P, then Q, P, therefore Q). But I might know that I can force the witness to admit that he did not see Q. I could then try to get him to acknowledge, or simply argue to the jury, that he thus could not have seen P (by modus tollens, if P, then Q; not Q; therefore, not P). In my closing, I could walk the jury through the inference patterns of modus ponens or modus tollens to show that the adversary's witness either (1) acknowledged propositions that are helpful to my case, (2) withdrew or qualified testimony that was harmful to my case, once confronted with the implications of his or her own prior testimony, or (3) denied these inescapable implications and thereby discredited him- or herself as self-contradictory (the classic ad hominem attack) and thus not worthy of belief.

After more than 33 years of conducting this exercise, I (along with my co-authors) decided to try a deeper analysis of the interplay between logic and cross examination, with some common-sense trial experience intermixed. It now seems to me that logical implications dictate the strategy of cross-examination, but that the questioning is a matter of rhetoric. No one cross-examines a witness using syllogisms per se. Rather, the effective trial advocate, through his or her questions, develops testimony and signals inferences that the jury will grasp, recall, and care about, and this presentation requires an exercise of judgment as to which facts will drive the most helpful inferences (and will not drive any harmful inferences) and, tactically, how best to signal the desired inferences effectively to the factfinder (without backfiring).

Aristotle tells us that the arguments proper to the trial advocate are enthymemes, which are the very "substance of rhetorical persuasion." Rhetoric, I.1 1354a14-15 in The Complete Works of Aristotle, Vol. II (J. Barns ed., The Rev. Oxford Trans., Princeton 2014). Enthymemes are sometimes described as incomplete syllogisms, but they are better defined as syllogisms suitable for rhetorical purposes, which does typically mean that they do not fully explicate the premises and conclusion of the argument. Rhetoric, I.2 1356b4-5 (defining the enthymeme as "a rhetorical deduction"); id. at II.22 1395b25-28 (arguing that certain "steps" in an argument should be omitted, so as not to "waste words in saying what is manifest"); see generally M.F. Burnyeat, Enthymeme: Aristotle on the Logic of Persuasion, Aristotle's Rhetoric 3-55 (David J. Furley & Alexander Nehamas eds., Princeton 2015).

There are at least five obvious reasons that arguments in oral advocacy are not stated in complete logical form: (1) it would be too time-consuming and belabor the obvious; (2) if complex, it would be taxing on the jury's attention and difficult to follow; (3) providing the animating premises of an inference, and allowing the jury to reason to the conclusion, might make a deeper impression, for "whatever strikes the mind of a juror as the result of his own observation and discovery makes always the strongest impression upon him, and the juror holds on to his own discovery with the greatest tenacity and often, possibly, to the exclusion of every other fact in the case." Francis L. Wellman, The Art of Cross-Examination 184 (4th ed. 1962); see also Rhetoric II.23 1400b29-33 (observing that "part of the pleasure we feel [in following deductions of which we can foresee the conclusion] is at our own intelligent anticipation"); (4) if the witness can see where you are going, he or she will try to avoid your conclusion by shaping his or her testimony; and (5) there are premises and possible inferences that might technically support your case, but that should not be emphasized because they might backfire—trigger inferences, associations, or emotional reactions from the judge or jury that undermine your case. It is this last point—determining which premises to elicit and from which witness, which inferences to articulate and when—that is one of the most prized skills of a trial lawyer. So there is an art to selecting the premises to elicit, the inferences to state expressly, the conclusions to try to force the witness to acknowledge. See Rhetoric, I.1 1355b9-11 (arguing that the "function" of rhetoric "is not simply to succeed in persuading, but rather to discover the persuasive facts in each case."); I.2 1355b 27-28 ("Rhetoric may be defined as the faculty of observing in any given case the available means of persuasion."). Two examples illustrate this point.

The first brief example is drawn from Francis Wellman's classic treatise, The Art of Cross-Examination, and concerns a case in which the defense lawyer (Wellman, himself) elicited on cross-examination only the minor premise of an argument he made to devastating effect in his closing. See Wellman, supra, Chapter 23 (The Bellevue Hospital Case). A reporter had managed to gain admission to Bellevue Hospital, for the purpose of writing a sensational story about what he saw inside an "insane asylum." Id. at 414. He later became the star witness in a manslaughter prosecution, in which the prosecutor alleged that the defendants had violently and repeatedly strangled a patient with a bedsheet while berating him because he refused to eat. Id. at 414-16. The reporter testified at trial that he was an eyewitness to the crime, that he had realized, as he watched, that "here, before [his] eyes, a strangle was going to be performed," id. at 415, and that he "stood by in that part of the room, pretending to take away the dishes[,] in order to see what was going on." Id. at 430. The defense lawyer, however, confronted the reporter with the fact that the reporter had failed to mention this alleged homicide in a newspaper story he wrote the very same night he left the asylum, and in an affidavit made two days later, and in a subsequent statement, and also in sworn testimony at a coroner's inquest. Id. at 426-30. The reporter maintained that he had forgotten the episode on these prior occasions, claiming that, because he had not known the man had died, "there was nothing to write about." Id. at 426. In closing (and not on cross), defense counsel made the following argument: "Can the human mind draw any other inference from these facts than that he never saw it—because he could not have forgotten it if he had ever seen it? If he never saw it, it did not take place. He was on the spot, sane, and watching everything that went on, for the very purpose of reporting it." Id. at 430-31. The lawyer did not pursue on cross either the major premise (if you saw it, you could not have forgotten it) or the conclusion (you never saw it) of the argument he presented so logically in his closing. But lest we conclude that this approach is generalizable, one ought to see that these were rather ideal facts for a defense lawyer and few prosecutors would put a witness on the stand who would so readily and utterly self-destruct (and who, in fact, refused to return the next day and fled the jurisdiction—surely the gold medal for a cross-examiner). Id. at 435.

Our second example is a hypothetical that we think is, while fictional, more realistic than Wellman's historical example. The fact that it does not lend itself to the crisp syllogistic presentation that was achieved in the Bellevue Hospital case illustrates, we think, the range of pressures and constraints affecting the rhetorical decisions of a trial advocate. Consider a white-collar trial, where the government has charged a number of executives with a complex fraud scheme. Assume that the government's case depends heavily on the testimony of a single cooperating witness (C), who tells the jury which defendants were part of the conspiracy to defraud, which meetings furthered the conspiracy, and which transactions were undertaken with a fraudulent purpose. Let's focus on one of those executive defendants (D) and assume that he was only tangentially involved in the events at issue, e.g., he was party to certain emails that incidentally connected him to key events. Assume further that the statements attributed to C in the material produced by the government pursuant to 18 U.S.C. §3500 (essentially, C's prior statements on subject matters inculpating D and others) do not directly implicate D in the fraudulent scheme or only mention D's alleged complicity as an afterthought. Nonetheless, in light of the 3500 material, D's trial lawyer expects that C's testimony, and the manner in which the government will likely question C on direct examination, will attribute to D culpable knowledge and involvement in the fraudulent scheme, even though none of the other documentary or testimonial evidence concretely supports such an allegation.

In reviewing the 3500 material, and preparing to cross-examine C and the other government witnesses, the defense lawyer will explore the implications of what C said and failed to say. The defense lawyer will try to cast doubt on C's attribution of illicit purpose to D and suggest other plausible explanations for D's alleged words and actions, all without drawing unnecessary attention to D's role in any of the allegedly illicit events. An implicit attack on C's recollection, understanding, and forthrightness might be more effective in this context than a full-frontal assault on the facts alleged. The lawyer could expose evidentiary and logical gaps by exploring the implications of C's testimony in order to establish that, for example, (1) although D might have been present for certain discussions or copied on certain emails, C had (erroneously) assumed D's complicity (despite the lack of "smoking gun" statements or acts) and that C had sometimes even assumed D's presence and participation in certain events, (2) C's understanding of the transactions in question was incomplete or incorrect, and (3) C erroneously (maybe even vindictively) lumped D in with others (who might be culpable) because of a close personal or business relationship between the co-defendants. In each case, the defense lawyer will rely upon an important but unstated major premise: If D was complicit, then D would have spoken in furtherance of the fraud at the meeting (versus merely being present) or written or at least responded to the damning emails (versus merely being copied); if C understood the fraudulent purpose of these transactions, then C would know why they were structured in one way and not another; if C perceives D to be close with co-defendants, then C is more likely to attribute the same knowledge and intent to D and the co-defendants.

But this last point is tricky because it could potentially implicate one or more of the co-defendants. Unlike in criminal cases involving a single actor or a discrete event, white-collar cases typically involve multiple individuals and numerous transactions. If it were a good strategy to point the finger at a co-defendant as an alternate target of blame, there would be several such targets sitting nearby. But defendants in white-collar cases are often extremely reluctant to point an accusing finger at a co-defendant. The reasons are complex—doubtless involving social and financial pressures—but one obvious reason is that it is better for all the defendants if the government is unable to prove that there was any conspiracy to defraud. It is likely a losing defense to argue that the witness's testimony is evidence that other co-defendants were part of a conspiracy to defraud, but that the testimony is consistent with D having had a different purpose and a different role in the same events.

The lawyer may or may not confront C with the implications of any concessions C makes on cross-examination. Indeed, the conventional wisdom is that a trial lawyer should not address the logical implications of a witness's statements during cross-examination. Conclusions are for closing. Rather, what the defense lawyer must draw out analytically in real time are the presuppositions of a witness's testimony, not the implications. But the lawyer will certainly argue to the jury that the absence of 3500 material (i.e., C's pretrial statements) regarding, e.g., D's presence, statements, or actions at a given meeting implies (by modus tollens) that D was not there, or that C never told the government pre-trial that D was there, or that C was embellishing a hazy and slight recollection regarding D when C testified that D was there. Likewise, if C lacks any of the other knowledge one would expect of someone who could reliably testify as to D's presence or illicit purpose, then the defense lawyer can argue that C's lack of such reasonably-expected other knowledge makes C's testimony about D's illicit purpose or knowledge unreliable or that C must have some other motive to testify against D (maybe one the defense lawyer can substantiate). None of this is the sort of crystal-clear syllogism that Wellman was able to produce in the Bellevue Hospital case. But it is nonetheless a serviceable enthymeme—reasonably clear, memorable, and commonsensical—that supports a not-guilty verdict in a trial where there is some evidence of guilt.

In conclusion, this discussion suggests two basic principles of cross-examination. First, during trial preparation, anticipate the witness's likely testimony, map the presuppositions and consequences, and identify the concessions you can elicit from the witness's own likely testimony that will support the inferences you want to argue in summation. Second, in light of your overall trial strategy—considering among other things your key themes (grounded in the elements of the claims and defenses), the body of trial evidence, the psychology and motivations of the witnesses, and the sensitivities and vulnerabilities of your client (and maybe the co-defendants)—identify those lines of cross-examination that are likely to yield memorable, emotionally compelling, and comprehensible admissions and inferences that support a judgment in your favor. The science of logic in the preparation; the art of rhetoric in the courtroom.

Milton L. Williams is a partner at Walden Macht & Haran, where he represents both criminal defendants and civil litigants in his trial practice. Brian D. Mogck is a partner at the firm, where he represents clients in complex civil and white-collar matters. Daniel A. Cohen is a senior counsel at the firm, specializing in commercial litigation.