Building an idea system requires many step-by-step processes: which data should be coded into information, which code to use, which relational bridge to activate, how to extrapolate secondary ideas from an initial one, and so on. These cognitive processes have a chain-like structure: the establishment of one chain enables one idea to be extended to another. An enormous, intricately structured chain-complex of ideas do not happen easily or perfectly; they are built and maintained by both routine and creative works.

This chapter is intended to capture several core dynamics. From an original idea, step-by-step idea-building processes could extend very far, connecting to many interlinked ideas. At the same time, the speed of operation could be very fast once these chains are established, enabling people to rapidly learn how to negotiate the entire structure. Missteps in one individual segment chain—even if they are “minor”—could create catastrophic chain reaction; in other words, things could go very wrong very fast, with very far-reaching consequences.

Two Types of Stepwise Inferencing

That ideas transfer through chain-by-chain “rational” processes has been theorized by Philosopher Charles Sanders Peirce, who posited deduction, induction, and abduction as three inferencing types.Footnote 1 For our purposes, I am simplifying and combining these three forms of inferencing into two: stepwise deduction and stepwise idea creation.

Stepwise deduction means that once a person takes an epistemic step, it can lead to a “logical” or a “plausible” next step. Deduction includes step-by-step calculation, weighing, evaluation, and comparisons.

Here is an illustrative example: if killing one person deliberately warrants the death penalty, despite justifiable circumstances, what about the act of killing multiple babies, even if one takes account of possible extenuating circumstances?Footnote 2 Killing a witch convicted of serial killing, all else being equal, is a rather rational, next-step conclusion following a previously established premise.

Likewise, if an old man who has confessed to having practiced witchcraft for decades but only remembers five accomplices—“all dead and all more than six years ago” according to the verdictFootnote 3—it would then be “logical” and “rational” to suspect that the person might be withholding information.

Stepwise idea creation means that once a person takes an epistemic point to be true, it can lead to the birth of a new idea based on the original one. People can do so by taking the cognitive steps of interpretation, guesswork (conjecture), speculation, extrapolation, or imagination.

We can consider how the establishment of this one fact generated a series of other ones connected to it: that having blood on one’s hands potentially could lead to God’s punishment. This idea had engendered the fear in judicial actors for being held responsible for the deadly prosecution of criminals. Attempting to deflect this responsibility, various judicial bodies in medieval Europe had devised “judicial ordeals” tied to various divinity tests. Couched in meticulous procedures and explicit pronouncements, these divinity tests invited God to adjudicate worldly affairs, and which implicitly defer the responsibility of killing to God. When England adopted judicial “firing squads” in the late nineteenth century, the governing body arranged the mixing of blank shells with real bullets, so the exact individual responsibility for killing was never determined.Footnote 4 These inventive ideas are next-step derivatives that have roots in an original one.

Step-by-Step, Chain-by-Chain Constructions

The extreme treatment of the accused during witch trials owed their justification to numerous works of stepwise constructions of a massive chain-complex of ideas. Some witch trials permitted the application of painful procedures far harsher than for normal crimes, even when the subjects were elderly. Furthermore, some practices seem to have been utterly unnecessary for prosecutions—such as stripping, shaving, and pricking.

The Bizarre, Brutal Acts Toward Magdalena Bollmann

Historian Lyndal Roper captures both the brutality and bizarre qualities in the case of Magdalena Bollmann of Marchtal in Germany in 1747; the suspect:

was tortured with thumbscrews, on the rack, and on the ‘bock,’ the bench on which she was stretched and whipped. She was stripped and shaved, and needles were inserted into areas around her genitals to see if she experienced sensation; they also hung her up and burnt her with a blessed Easter candle (a holy object to shield any magic) near her nose and under her toes.Footnote 5

If we are to deconstruct the step-by-step, chain-by-chain constructions that led the idea system to warrant such cruel and unusual treatments, we would first find the chain to be very long.

Long Chain Extensions: Unto the Very Far

In the chain of empirical codification, when deconstructed, each idea—which we can visualize as an individual link—only has a limited amount of potential meanings that it can develop; its actual force is also bound by the strength of evidence it has. Linking and bonding ideas structurally gives each idea new potentials, while bolstering the integrity of the overall idea system that supports certain meta-units and abstract constructs. Each chain unit can do much more in a system of division of labor.

One can also envision that a chain-complex has certain loose sections. Suppose two links are disconnected, then such disconnection may exist as a “gap” in our thinking. In such a situation, an intermediate (or intermediary) idea—like a few new, cheap plastic links, or even a whole new section made of metal—can help to bridge the parts to preserve the shape of the whole complex. At an even more microscopic level, suppose the metal in one particular link is separating, filler metal—akin to small pieces of information—could also serve bonding functions.

The causal power of intermediary ideas. Suppose a few witch-suspects are observed to display unusual masturbatory behaviors in a jail or in a nunnery, the acts by themselves could only indicate so much about witchcraft. No intermediary ideas are present between masturbation and witchcraft. The act itself may indicate something unusual, but the referential distance to witchcraft would be too far-fetched to be reasonable. A suitable, hypothetical example (though not so hypothetical as I will show) for an intermediate idea would be “demonic possession.” Individuals can then make this connection: that subjects masturbate unusually because of demonic possession. Each intermediate point may need more substantiation—having its own mass and weight, so to speak—in order for it to be a fastened point in itself. Therefore, “evidence” about symptoms of demonic possession might be solicited, making it easy for a subject to imagine a direct connection between demonic possession and masturbation (or sex, or social violation, more generally) on the one hand, and demonic possession and witchcraft on the other.

The more intermediate ideas a system contains, the more likely these ideas—even when held in reserve—could be candidates to connect ideas that otherwise would hit a dead end to their development. Sometimes individuals artfully connected disjointed ideas ingeniously; the corpus of the witch-hunt idea system contained many intermediate ideas that could fostered such artful actions. However, certain institutions practically forged intermediate actions to make and fasten a variety of linkages particular to specific cases.

Torture was a great example. Many prisoners could not withstand torture and had admitted to something petty at first that was unrelated to witchcraft—for example, they harbored ill wishes toward a neighbor, or they consulted a magical practitioner once upon a time. But that first admission was an admission to character flaws, which delegitimized their credibility. The inadvertent creation of that intermediate idea (character flaw) legitimized new actions—the falling trust in the subject’s previous and subsequent words as well as the need for them to divulge more detailed information that had not yet been elicited. Upon further questioning or torture, many suspects crossed an important line to admit to having aided the Devil or signed a pact. They might say that they were unwilling in their action—such as being beaten or tricked by the Devil, and they might not even know that their actions would have led to deaths, illnesses, and other misfortunes. Such admissions would reveal further character flaws. If they admitted to appalling deeds, even unwilling participation in them would be extremely damning. At that point, they had little credibility at their disposal other than to provide a satisfying account—the “true” and “frank” confession. After all, if a person is so evil as to eat babies, for example, then why should we trust any testimony they could provide in favor of their acquittal? This is one way in which a chain reaction could occur within a case, where an intermediate idea of character flaws finds new reinforcement with each step, and the fall from grace—one link more each time—engendered a further link of intermediate interrogation that yielded information justifying the harsh elicitation methods. The first admission of a character flaw, in other words, enabled a long chain of stepwise extension and relevant stepwise actions.

The implication of this chain construction goes beyond individual cases. On a broader scale, an ever-escalating abuse of torture could be justified. In other words, if escalated torture is shown to have worked to make a witch confess, then the need for escalation becomes an intermediary idea in itself. There are some subtleties involved, however.

For ordinary crimes such as murder or theft, even if they were considered despicable, punishment was rather mild and torture ended relatively quickly compared to the many cases of witch hunts. Brutally extracted confessions run the risk of being considered barbarous by the public. In ordinary people’s understanding, more force certainly would extract more confessed details about a case. It is natural that people would be automatically skeptical if they see a direct connection between the application of force and the confession. Not only would they question the subject’s account, they would also question the legitimacy of the judicial body.

Comparing diabolical to heretical crimes, a heretic still maintained a mortal identity, and his or her power was understood to be quite limited. But a diabolic crime implied not only the possibility of great social damage—which in itself might legitimize more torture due to a greater urgency—but more importantly the power of supernatural entities was involved. The straightforward relationship between force and confession was obfuscated by sensational, intermediary ideas generated about the supernatural at every step. More torture yielded more confessions; more information about the unknown was produced, which, in turn, called for more intermediate actions. In this way, the appearance of barbarity was often masked by the veil of supernatural dramas.

Escalating torture often incites the intermediate action to grow the pool of information providers. Both ordinary and extraordinary crimes may involve suspects betraying their conscience and yielding a false confession, but in the case of ordinary crime, to do so would not prompt one to invent or support some powerful intermediary ideas. The conviction of a home burglary, for example, does not subsequently implicate more than several other individual collaborators at most. This dynamic played out very differently in mass witch trials, where suspects seemed to multiply at great speed with the application of torture (or the threat thereof), and with each confession, new intermediary ideas were created. This was fueled by the widely held belief that witches often behaved in groups, while the more insidious kind would repeatedly commit acts of diabolical treachery. Table 5.1 captures and models this self-perpetuating dynamic.

Table 5.1 Stepwise dynamics in prosecutions of extraordinary crimes
Full size table

A particularly powerful example of an intermediate idea worthy of special mention is the breaking point of witches—the idea that witches needed to be brought to a “breaking point” for them to confess.Footnote 6 A witch’s breaking point was different from that of an ordinary person’s because of possible diabolic arts or the intervention of the Devil himself. The details that give rise to this idea come partly from textual interpretations, but partly from witch confessors who were tortured. Consequently, the details yielded under torture could actively substantiate theories that justified further torture.

Witch confessors not only had furnished the means upon which witches could commit crimes but also withstand torture! If a person refused to confess to an ordinary crime, the resistance might indicate actual innocence or the person’s personal stubbornness and ability to withstand pain. But a competing intermediary idea was that criminal tricks could be used to mitigate the effects of torture.

There were slight variants of these conceptions. One version was that the Devil could help a subject to withstand torture—one alleged means included making subjects insensitive to pain; another was to make them “fall asleep” during torture. Yet another was that the Devil could cause witches to make certain statements during interrogation, and even recant their earlier confessions.

In the Compendium Maleficarum (1608), Francesco Maria Guazzo selected several cases borrowed from Martin Del Rio, Nider, Remy, and a book author and legal consultant called Loys Charondas le Caron (author of Antichrist Unmasked) to conclude that witches under the influence of the Devil may not feel pain during torture.Footnote 7 In one case he cited:

Near Amiens in 1599 a girl witch was imprisoned, who felt nothing when her feet were cruelly burned or when she was heavily scourged, until at the suggestion of a priest they hung about her neck a waxen image of the Blessed Lamb. Then, by virtue of the sacred amulet, the wiles and guile of the devil were defeated, and she began to feel the force of pain. Therefore it is clear that this indifference to torture, which even Tostado (In Genesim. XIII) recognizes, springs from no physical cause, but is due to the devil’s work.Footnote 8

The implication is very important. Whereas, for an ordinary criminal, successfully withstanding torture could be a sign of innocence, in this conception it could indicate supernatural intervention by the Devil on behalf of his minion!

Another intermediary idea along similar lines is that witches could make themselves insensitive to pain—or to faint, fall asleep, act delirious, and so forth—by using diabolic arts or even using natural means. The consumption of drugs prior to torture was frequently mentioned by demonologists. Henry Boguet argues that such substances were used by “more cunning witches”Footnote 9—presumably in comparison to those less cunning witches who did not use such tricks and confessed frankly. On the one hand, the drugs could be made of something diabolical, such as an ointment that constituted a “Spell of Silence”;Footnote 10 on the other hand, it could be something entirely natural—a drug made of a mix of common herbs, or “even salt dissolved in clear water”—that could help witches withstand torture. The witches could be acquiring any preemptive substances from prison guards who might be selling them to supplement their meager wages.Footnote 11 Witches did not always have to use mystical means to do harm; they could also use poisonous herbs—a natural means—to kill animals and people. It might be considered a “poisonous magic.”Footnote 12 But witches were associated with being skilled at making poison and compelled to use their arts for evil purposes. For such witches, the breaking point theory might not strictly apply. But raising the threshold of torture becomes more legitimate, as harsh torture had little effect on them anyway.

This leaves the question of how to overcome these diabolic arts of resistance?

The dominant theological opinion was that witch’s resistance to torture could be overcome. Past cases gave evidence that torture could do the task, by first creating a crisis in the witch. At that point, the power of the Devil or the diabolic art would be overcome. A floodgate of confessional statements would follow soon after the “breaking point” was reached.

Indeed, if some accused parties withstood nearly the most serious of forms of torture available and then yielded an elaborate confession, they would almost be handing witch believers an exquisite gift. These high-quality subjects reinforced the ideas regarding witches’ breaking points, and the harsher torture needed to overcome it. If the accused had also had previously enjoyed a respectable reputation, and had yielded an extremely persuasive defense initially, then the case could show that even such respectable people and persuasive defenses could not be trusted. This is particularly the case if their final confession accounts turned out to be well verified and extremely detailed, bearing remarkable convergence with other testimonies.

In fact, even the cases wherein an accused withstands torture and is set free could feed into the intermediary idea. Consider Maria Holl of Nördlingen, Germany, who was set free after a lengthy torture—after more than sixty applications.Footnote 13 One still could say that truly honorable characters could withstand exceptional torture, with her as a shining example.

The same goes for the accused who “voluntarily confessed without torture.” While the phrase might just mean they do so in a court after being tortured, there were plenty of others who voluntarily confessed upon being threatened with torture. These subjects also strongly feed into the theory of witchcraft, by the information and ideas they provide. It is a different kind of horror altogether from supporting a theory of the breaking point.

Shall the suspects commit suicide during detention, instead of being brought to confession, an intermediary preemptive idea was often evoked by demonologists. Per these demonologists, the Devil often tries convincing witches to commit suicide, foretelling them the suffering ahead, and by doing so they make the witches forsake God and their own salvation.Footnote 14 If the suspects died during torture, sometimes the outcome was attributed to the Devil’s deeds—perhaps “performed by the infamous Foe [i.e., Satan] to prevent them [the guilty witches] from receiving Grace from God through Sacramental Confession.”Footnote 15

In sum, all these intermediary ideas lurking quietly in the background could be evoked strategically depending on the need to extend or improve the chain-complex. Once put in motion, they do more than inspire cognitive actions. They inspire and justify a broad range of intermediate actions—encompassing various physical and institutional activities that are possibly as far-reaching as imagination itself.

Portal-Like Chain Extensions: Unto the Very Fast

A big idea complex might seem to take a long time to navigate. But this is not necessarily the case.

We think swiftly, and we think slowly. When we think quickly, it could be because the problems are so simple that they hardly warrant any cognitive labor; or that we are such lazy and hasty thinkers that we simply jump to conclusions. Yet, there is another scenario. After mastering a certain difficult craft—to the smallest detail—our knowledge becomes so clear and thorough, many difficult connections become so “self-evident,” then the speed of processing difficult tasks becomes drastically less. Like a weightlifter trained to lift at 400 pounds asked to lift at 100 pounds, the easier tasks could be processed swiftly because they hardly take any work anymore. In the vast extension of an idea complex, speed may correspond to our fluency with the idea system, in addition to our own mental power. A combination of dexterity, techniques, and practices in codification contributes to our speediness in idea construction.

Speed is also tied to whether cognitive portals are cultivated.

Cognitive portals are a means akin to cultivated tunnels that enable the speedy transference from one point of origin to the point of destination, with minimum impedance.Footnote 16 In scientific fiction, portals often enable subjects to suddenly transfer across space or time. Less magically, we can also imagine a secret pathway cultivated in a maze that enabled actors to suddenly transport from one place to another. Applying this metaphor to cognition, I mean pathways that seemingly enable people to exercise a “leap” in their thinking—skipping many steps—a process that is mystifying to outsiders. These pathways could also experience blockage by unpredictable changes or outside forces.

Cognitive portals can be located within individuals, without them being widely shared. These are private cognitive portals. Public cognitive portals are located in collective or public settings, understood and utilized by a number of listeners for collective, communicative, or social purposes. Then, institutional cognitive portals are built into institutions, serving institutional purposes. They do not have to be known in the wider public, nor are they only expressed in mental exercises. The flow of increasing forms of torture to extract more ideas that justified escalated forms of torture represented the existence of such an institutionalized cognitive portal.

In sum, Europe’s witch hunts may seem to be driven by hasty actors who exercised imprecise, fast-and-loose association and faulty lines of reasoning filled with chasms and gaps. These possible actors notwithstanding, the phenomenon should be linked to those of more serious thinkers—notably demonologists—who practiced incredibly incremental, meticulous theorizing and information-gathering. Sometimes, much individual thinking had already been done by others. The hasty thinkers could afford to be so hasty because pathways had been cultivated for them, with exemplary role models serving as their guides. Discursive and institutional protocols could be disseminated to the public, making certain lines of thinking very easy to use and to follow. Their “jumping to a conclusion” and making “cognitive leaps” were not only systematically taught, but they were justified and protected by select manipulators of the idea system who cultivated various pathways and portals.

Extending and Multiplying Errors: Unto the Very Wrong

Gaps, errors, and misinformation inevitably exist in human reasoning. There are certain conventions of acceptability to flaws in reasoning processes—and in codification activities. Our interest here is how idea systems managed to create ideas that became extremely “wrong”—as in distorted and erroneous. Building on mechanical metaphors, the structural explanation in this chapter is admittedly partial. It focuses on how flaws do not necessarily emerge as a single phenomenon; flaws could transfer and multiply in successive steps of codification; dynamic effects could occur by way of subtle changes, as ideas are passed along a chain of compounding inferences. Figure 5.1 sketches several dynamics of distortion emanating from reasoning flaws.

Fig. 5.1
An illustration of 3 scenarios depicts the metaphors of sound amplification, prism, and the game of telephone with the help of the chain of nodes.

Three linear distortion dynamics

Full size image

In these graphic representations, each node on the left side represents a point where an idea begins. As it moves rightward, it experiences a point of transition or transformation—which can take the form of a re-representation, restatement, translation, simplification, rearticulation, or various forms of stepwise deduction and idea creation. A distortion caused by reasoning flaws could be introduced at each step: such as a misapplication of a text, a slight overstatement based on empirical patterns, or a small conflation of concepts. I use changes in shape, shade, and direction to represent such distortions.

Scenario A uses a sound amplification metaphor. If we have five nodes and four processes, and if an 8.3% (of 100) subtraction occurs every time a “step” is made in inferencing, then 66.8% remains at the end of the chain. If we are merely amplifying a sound, this 66.8% probably is not too damaging. And a device at the end of the chain that can reamplify the sound with additional gains can “correct” the distortion (caused by lowered sound volume).

Scenario B uses a prism metaphor that “distorts” the direction of the light by about 15 degrees—approximately 8.3% out of 180 degrees.Footnote 17 A totality of 60 degrees of distortion occurs after four successive steps. But two compounding factors emerge. First, we can imagine that at each new step, the new distortion seems to be slight—with about 91.7% of accuracy—but each slight distortion “adds” to a prior distortion. Second, the final directionality is at 120 degrees, rather than the original 180-degree point of the compass. While the directionality did not switch completely from east to south, it became closer to the south than to the east. And, every step that is “correct” afterward still adds to the error, a dynamic that is different from Scenario A.

Scenario C uses a metaphor of children’s game of telephone. In contrast to Scenario A, we can imagine several people lining up and passing a message in the game of telephone, but one mischievous person at the third node deliberately changes the message altogether. The final accuracy would be at 0%. In fact, the outcome of 0% would be the same even when the chain is extended to twenty or more people passing messages, as long as only one person in the chain is mischievous. One hundred percent of distortion would be fully passed on, even if most people in the chain are 100% correct. We can relate the challenge of telephone to everyday organization. Often, the original intention and objective of higher-level decisions are simplified, lost, translated, and remodified as they pass through multiple meetings, committees, subcommittees, and interpersonal conversations to the point they can become a different animal by the time the “message” reaches the bottom level, and vice versa.

The Chain-Complex that Entrapped Magdalena Bollmann

Let us revisit the travesty of extreme treatments toward witches described by Roper, reflecting on how they are produced through many steps of inferencing along the structure of a chain-complex. Some step-by-step mechanisms that produced brutality have already been explored; we can now engage the step-by-step codification processes that cause the utmost bizarre treatments. The quality of certain key intermediate ideas greatly influenced the outcome. Because those key intermediate ideas were so bizarre and taken to be possibly true, so too were the “rational,” stepwise responses invented toward verifying and handling them.

One intermediary idea was the Devil’s marks—a coded thing in the idea system. This concept triggers many legitimate questions down the chain. If it is known that the Devil sometimes leaves physical marks on the body that could be used as empirical evidence, would it not make sense, as a next step, to search for them? And how do we search for them? What equipment would we use? What did the Bible indicate about the matter, and did past cases reveal anything?

Following these lines of thinking, shaving of witches’ bodies emerged as a “reasonable” means of investigation: the Devil’s marks could not be discovered unless the subjects were shaved all over. And once a suspect mark is identified, it could not be truly verified unless it was tested with needles. This is the reason for Roper’s account of Magdalena Bollmann, who was subjected to needles to test for sensitivity, and the same reason Urban Grandier also had a needle “buried” into his body at least an inch. And in Bollmann’s case, because weird marks were found around her genitals, the ultimate “conclusion” after multiple steps was to have needles inserted around her genitals, after being stripped and shaved.

The details from past confessions reinforced the justification for shaving—to remove the “spell of silence” previously mentioned.Footnote 18 As to why Bollmann was burned with a blessed Easter candle, the intermediary idea is that the candle, as a holy object, can counter magic; torturers who had burned it “near her nose and under her toes” had found abnormal responses.Footnote 19

Refraining from overspeculation, the general pattern is that one outlandish idea seems to, instead of existing in a static state, breed a complementary one, possibly followed by more. The chain of outlandish ideas that gradually emerges reminds us of the three aforementioned metaphors of distortion—where even minor distortions, occurring in particular manner within the chain, can drive an idea’s development to unpredicted trajectories.

Initially, some of these treatments were posited merely as experiments, which would test out intermediate ideas. But repeated experiments yielded findings, which then gave a degree of substantiated proof to those ideas, solidifying them into fact-like constructs. The Malleus Maleficarum confidently recommended a method of using wax and holy objects: “put a drop of Blessed Wax in a chalice or cup of Holy Water and give it to them to drink three times on an empty stomach in the name of the Most Holy Trinity. By the Grace of God we have removed the sorcery of silence from many in this way.”Footnote 20

The minutiae of selecting the appropriate code and information could matter a great deal. The specific properties of things can evoke corresponding, specific treatments, and bizarre intermediary ideas can inevitably beget eccentric actions and ideas. Forcing a subject to drink slightly waxed water was not as damaging as burning them with a candle. Unfortunately for onlookers of history, Kramer and Sprenger were exceptionally creative in generating and then justifying eccentric ideas. For example, in the Malleus Maleficarum, they advised that the witches, upon arrest, be hoisted off the ground and put in a basket. Because from past experience and confessions they learned that: (A) “Quite often they have lost the sorcery of silence through such a method of arrest”—that is to say, the subjects started to confess when they did so. (B) Many witches about to be burned requested that their feet touch the ground. (C) When inquired further why they wished their feet to touch the ground, they answered that if they had done so, they would have freed themselves, killing many people with bolts of lightning!Footnote 21

The intermediary idea of demonic possession leads to the idea of exorcism as a handling method. And once we accept the idea of exorcism, a plethora of bizarre treatments can be defined as appropriate. Exorcism was a means to deal with nonhuman entities. The normal boundaries associated with dealing with humans were expanded as the definition was transformed to the realm of the supernatural. This could include drinking weirdly mixed alcohol, putting a mixture of “sulfur seasoned with galbanum, asafoetida, and rue” on live coals and holding it under a person’s nose, or whipping a bewitched person’s body part where the evil spirit is “caught and pinned down,” while spraying “lustral water” at it.Footnote 22

The recommended treatments do not have to induce any pain. For example, if the idea says that boiling “victims’” urine would cause a witch to exhibit signs of suffering, it would simply lead people to watch the witches’ physical expressions closely while doing so.Footnote 23

In summary, Magdalena Bollmann was entrapped in an idea system that had already developed into a vast chain-complex, containing many intermediate ideas. Such chain-complex was not built overnight or all at once, but through many years of accumulation across time and space, on elaborate treatises and the painful experiences of many subjects. The people and institution who led Bollmann into extreme, bizarre torture most likely did not only employ “reasonable” steps in thinking and codification. They also did so very quickly, as they were enabled by various cognitive portals that had been built into the idea system. The far-reaching, thundering “reasonable” steps people had walked include erroneous missteps. Such successive missteps were not entirely determined by walkers who chose their routes, but they were co-determined by those many actors who had cultivated those intricately connected pathways as possible choices.