Huntington, Janowitz, or None of the Above? Conceptions of Professional Norms by Future Army Officers

Through a survey of 1,470 cadets at the U.S. Military Academy, we measure the strength of civil-military norms in the midst of their socialization to civil-military values. In addition, we use an embedded list experiment to measure true preferences to controversial questions in civil-military affairs, such as competing loyalties to democratic governance. To analyze cadet adherence to norms we first articulate a framework of various conceptions of professionalism. While cadets do not exhibit adherence to one particular framework, we find limited adherence to a purist Huntington model, and more of a conflicted reading of Huntington, as well as selectivity and superficiality in their commitment to norms, largely driven by their own partisan preferences. Lastly, we find concerning evidence that a significant number of cadets prioritize following orders over upholding democratic traditions, justifying Janowitz’s concerns that a military separate from society could breed attitudes among military personnel antithetical to democracy.

politics. We also include a list experiment in our survey to explore how much attachment to democratic norms are internalized as part of conceptions of professionalism. With the latter, we seek to understand how much soon-to-be commissioned officers prioritize a commitment to uphold the country's democratic traditions in their military service.
To guide our assessment, we outline several alternative frameworks or expectations about military professionalism, with the aim of analyzing which, if any of these approaches are reflected in survey responses. We intend these as heuristics, or organizing principles, to help guide our assessment of the cadets' survey responses. We characterize these three frameworks as: purist Huntingtonian norms; conflicted Huntingtonian norms; and one based on a selective adherence to norms.
We find that, while no single framework aligns fully with survey responses, there is only marginal evidence for the purist Huntington model, especially given the embrace of partisan behavior and other attitudes evinced by cadets. In contrast, we find some evidence of elements of a more conflicted Huntington model-one that proposes that Huntington supports contradictory beliefs and attitudes about political activity and civilian control-in cadets' responses. We also see evidence of some selectivity and superficiality in their adherence to norms, such that cadet responses are heavily predicated on their own partisan beliefs. Most alarming, in our list experiment, we find some justification for Janowitz's fear that professionalism, especially when premised on the military's separation from society, could breed attitudes among military personnel antithetical to democracy. This paper proceeds as follows: first, we outline in greater depth the four conceptions of professionalism we anticipate cadets might exhibit. Second, we detail the methodology we used and the parameters of our survey research, including the design of a list experiment we employed, aimed at better probing the true attitudes of cadets on sensitive matters, such as when upholding democratic norms collides with following orders. Next, we provide the findings from our survey of cadet attitudes and how well they align to the four models of professionalism we introduced. We close with a discussion of implications for the adherence to and understanding of professional norms within the officer corps.

Frameworks of Professionalism
Military professionalism at its most fundamental is usually associated with several key properties: the cultivation of expertise, the development of an "organic" ethos and corporate identity; on-going education; and an ethos of respect and responsibility for society. 2 It is also sometimes used to distinguish conscript armies, or those built from citizen-soldiers, from those maintained primarily, if not exclusively, from largely self-selected military personnel led by career military officers. While these basic attributes are often cited in definitions, scholars more broadly conceptualize the norms exhibited by military professionals differently; these approaches diverge in their expectations of the character of officers' beliefs, values and behaviors. These alternative conceptions of professionalism provide a template for assessing cadet attitudes. Each suggests that cadets will exhibit a distinctive pattern of beliefs. Below we introduce each approach and then discuss some of its implications for military professionalism.
The first is Samuel Huntington's influential objective control model, which we term the purist Huntington model. This model takes at face-value Huntington propositions about what should be observed in the behaviors and mindsets of officers who are military professionals. Many contend that objective control has deeply influenced military culture (whether or not individuals expressly identify that culture with Huntington). 3 Objective control, in turn, has particular behavioral and attitudinal implications. 4 Most prominent is Huntington's contention that professionalism requires a separation of spheres and that officers will focus primarily on cultivating their technical expertise in military affairs. Military and political leaders work in tandem in their respective spheres free from 2 According to the publication, Army Doctrine Reference Publication 1, The Army Profession, professionalism encompasses several tenets, including a commitment to and relations of trust with society, the acquisition of expertise, autonomy and self-regulation, and stewardship. Academic research by sociologists on military professionalism stresses other, related values, such as corporateness, responsibility, and expertise. 3  interference from the other. By extension, Huntington also implies a modal form of military advice to civilians that is highly transactional, in which civilians deliver guidance and military leaders supply options.
Huntington also requires and anticipates military professionals will remain apolitical; they will abstain from engagement in politics, leaving the latter to civilians. Note that Huntington's proscription against engagement in political behavior goes well beyond partisan activity and entails an intellectual distance or disinclination to engage with how political factors might bear on the efficacy of strategy or conduct of military operations. 5 In sum, a purist Huntington model would anticipate that cadets would endorse a clear separation of spheres and a belief that they should focus exclusively on cultivation of expertise in the "management of violence," while civilians exclusively make political decisions with respect to the use of force. Hence, cadets adhering to those norms would reject the idea that the military works on equal footing with civilians in advisory processes and believe it is not their job to worry about domestic political support for the war. They would also disavow engagement in any and all partisan activity. We would, for example, expect to see reticence to engage in political speech on social media. 6  the beliefs and behavior of military officers. 8 Its implications for military officer mindsets and beliefs is more complex and different than Huntington expressly posits. This represents a conflicted Huntingtonian normative framework. In other words, it allows for officers to agree with some tenets of Huntingtonian thinking, especially those requiring compliance with civilian orders and decisions, the maintenance of military autonomy and formal and explicit rejection of engagement with politics, while also evincing more conflicted views when it comes to actual behaviors and opinions about partisan activity and towards the practice of civilian oversight. Specifically, in the conflicted model, officers should endorse measures that protect their autonomy and ostensible subservience to civilian authority. Yet, as the result of "blind spots" they may also be more willing to resist civilian incursions into their sphere and to engage in partisan activity. In addition, contrary to what Huntington argues, his version of professionalism can create attitudes corrosive to civilian control, including doubts or at times even contempt for civilian leadership and suspicion of civilian motives and capabilities in exercising oversight. We would expect to see this reflected in cadet's responses if some of the perversities of Huntington's conception of professionalism are affecting their views.
In addition, the conflicted model suggests that objective control should enable the development of negative attitudes toward society. Huntington argued that military personnel exhibited a distinctive (and monolithic) conservative mindset at odds with liberal societal culture, such as that found in the United States. More controversially, he argued that the former was superior to the latter and that American society should try to emulate the military's conservative values and mindset. If this model shapes cadet beliefs, officers should exhibit negative, if not condescending, attitudes toward society.
A third argument is that socialization to norms of professionalism is only selectively observed. Military personnel have multiple identities (as do all individuals) and those other identities interact with their adherence to norms, such as they abide them only when they align with other key values or principles. These other identities might be based on gender, race or partisan affiliation.
Professionalism is insufficiently socialized, such that military identity remains subordinate to other societal values. When robust, professionalism should override other inclinations based on alternative dimensions of identity, such as partisan identification. As Sam Canter has recently put it.
"Professionalism is not creating the illusion that a general is a mindless automaton. Professionalism is the ability to put those feelings aside-to tamp them down as deep as they will go-and honorably serve the Constitution and the duly elected officials of this nation to the best of one's ability, regardless of political affiliation or outlook." 9 In other words, professionalism should suppress other identities, such that cadets identify most strongly with the values of their profession in answering questions about their future career as officers.
In contrast, if adherence is selective, we would anticipate that, while cadets do express views consistent with tenets of professionalism, those views are only limitedly informative of their attitudes and beliefs. Here we would expect that demographic factors and those not related to military service might better predict survey responses. Other attitudes override professionalism principles, such as partisan identification. 10 Consequently, we would expect to see divergences and systematic patterns in the normative frameworks or attributes endorsed by cadets according to their partisan or social background; responses reflect systematic patterns, but the biggest predictor of how one responds is partisan identification. For example, support for "water cooler" talk (informally talking about politics at work) would fall upon partisan lines, with those who view their partisan  These expectations are summarized in Figure 1. Observe that several of the predictions overlap. As noted above, we intend this table to provide a heuristic or a means of organizing expectations about how attitudes might conform, not as a predictive tool.

Research Design
In In addition to this descriptive battery of questions, we also included a rank-choice question, asking respondents to sort different potential attributes of "professionalism," as they understand it, from most to least important. Finally, the survey instrument includes an experimental portion, in the form of a list (or "item-count") experiment. These types of survey items have historically been used to assess respondent attitudes on controversial subjects, such as racism. We employ this technique to evaluate whether these future military officers would resist civilian orders that manifestly damaged the country's democratic institutions -and to what extent they potentially falsify those preferences.

Sample
Our sample was drawn from cadets at the United States Military Academy from December 2019 to January 2020, fielded in two waves to students enrolled in the introductory classes in both American politics and international relations. The survey's respondents constituted an opt-in panel, who received the survey online through the platform Qualtrics. Table 1 shows demographic and covariate balance statistics for the sample across both conditions of the experimental portion. In addition to a high response rate from the respondents, our sampling was able to exploit the underlying cross-section of the West Point cohorts, which are intentionally drawn from all Congressional districts, providing a representative population for our sample.
We should note that we do not expect that cadets have fully formed conceptions of military professionalism this early in their careers. Nonetheless, given the socialization that occurs among cadets through their West Point experience, we would expect them to reflect some ideas about the meaning of the profession and their roles within it. Some of these ideas are also likely informed by the experiences of their family members (especially given that many come from families with relatives who have served) and larger cultural conceptions of civil-military relations. For example, there are indications that many Americans are attached to Huntington's separation of spheres idea, believing the military should essentially run the wars, 11 and that they believe military experience makes for better leadership across a broad range of government roles. 12 We are interested to see what attitudes cadets exhibit, even while recognizing that they are early in their socialization process. civilian leaders function in an "unequal dialogue," where despite whatever technical expertise the military brings, fundamentally, they are subordinate to civilian authority and never operate on equal footing -a formula that the cadets we surveyed would likely protest. 13 This answer is suggestive that cadets reject the purist Huntington model, in which they would abstain from engaging in discussion with civilians about politics and policy.  Similarly, cadets we surveyed also indicated strong support for retired general and flag officers to play important roles in politics and government. Respondents believed 5-to-1 that more retired general and flag officers serving in the cabinet was "good for the country" (50 percent agreeing compared to 11 percent disagreeing). 16  third of cadets viewed discussions that occur in the barracks, office, or foxhole to be private in nature -even though they occur in uniform and on duty -and even when those discussions turn towards politics. This receptivity to partisan behavior is suggestive that either norms are not well socialized, or of the blind spots associated with Huntington's objective control.   Specific differences emerge between partisan subgroups within this sample of future military officers, particularly on questions pertaining partisanship and civil-military propriety. One way we can analyze underlying trends like this is through Principal Component Analysis (PCA), which is a form of dimensionality-reduction designed to uncover patterns and correlation in complex multivariable datasets. Our descriptive battery contained many different questions, but some of these are likely to be correlated with one another; as a result, reducing the dimensionality of the answers to only a few principal components can help us understand the sources of variation, as well as how answers to particular questions correlate with one another. Figure 4 reveals some of the most important results of PCA in a graph called a biplot.

Civil-Military Relations Norms
First, respondent answers to thematically-similar questions correlated along logical lines. The factor loading of a particular question (depicted a vector from the origin) reveals how the principal component is related to an answer for that item. For example, the vectors for approval of retired partisan commentary and retired critical comments load heavily onto PC1; the higher the PC1 value for a score is, the higher the approval for these comments. But these two vectors are also very close in magnitude and direction, indicating answers to them amongst our respondents were highly correlated. This makes sense, given the similar theme of the questions. The same was true for answers on partisan speeches and campaign speeches in front of the military and discussion of politics at work and social media.
Second, graphically depicting the subgroups onto the biplot also tells us how positive loadings and partisan identity are linked and correlated. In our sample, positive answers to questions on whether most of the military "shares my beliefs," on whether one party performs better at national security, and whether retired GOFOs in the cabinet was good for the country, are tightly correlated, loading heavily on PC1 and PC2. However, the direction of these vectors points towards the cluster of observations that identify as Republicans, and away from the cluster of Democrats.
This makes sense given the wide separation among partisans on these questions, while potentially revealing a more partisan thought process on these questions, orthogonal to other normative questions like talking about politics at work. Perhaps more importantly, the joint belief that (1) the

military shares Republican views, (2) Republicans do better on military issues, and (3) retired
GOFOs in government is a positive, potentially indicate more than partisan bias, but a belief in the military as extension of a conservative partisan constituency. This pattern seems to be especially apparent among cadets that identify as Republicans. In other words, those who are Republicans are more likely to see the military as their co-partisan and therefore to support greater influence in government, even beyond the national security domain. This partisan rationalization is consistent with the selective adherence normative framework, in which cadets' partisan identity is in tension with or overrides any principled beliefs about the military role as confined to cultivating expertise in its specific area of professional responsibility.
Other partisan divisions were less indicative of partisan rationalization and more of a commentary on the partisan make-up of the cadets we surveyed. Cadets who self-identified as Democrats were almost four-times less likely to say that most people they knew in the military shared their views (10 percent of Democrats compared to 36 percent of Republicans). This is consistent with past survey research throughout the All-Volunteer Force era, which has found Army officers, to include cadets and junior officers, who self-identify as Democrats to be a minority in the officer corps. 22 The fact that a majority of cadet respondents could even determine whether or not members of the military shared their partisan attitudes in the first place is noteworthy, as it suggests partisanship within the ranks is observable, if not known. Alternatively, it suggests perhaps that cadets perceive the military to be partisan with a gloss of nonpartisanship, rather than a place where the nonpartisan ethic of military professionalism is deeply socialized and suppresses partisan affinity. If professional norms of nonpartisanship were overriding partisan identities, we might expect more uncertainty or neutral responses in answers to this question.
Democrats were also four-times less likely to believe that one party makes better decisions about national security (14 percent of Democrats compared to 54 percent of Republicans). This finding carries real implications for healthy civil-military relations, especially as these cadets advance throughout their careers as Army officers. If officers who self-identify as Republican-who constitute a majority in the officer corps-believe one party is better at national security than the other, will those officers fully implement policies directed by their Democrat civilian overseers?

Aspects of Professionalism
Each of the aforementioned batteries aimed to gauge cadets' views civil-military relations norms, but we also wanted to explicitly ask them what they thought were the most defining characteristics of being part of a profession. Figure   These findings too are suggestive that purist Huntingtonian norms explain cadets' beliefs.
He argued that professionalism would impart a sense of value for expertise and respect for cultivating it, and this seems to be clearly indicated in their ranked ordered responses. Yet, the fact that apolitical remained so low on the cadet's responses is contrary to the purist framework. It might be possible that cadets are simply unaware that they should remain apolitical, but their responses to other questions, such as that on social media behavior, belie such an argument. Rather, it appears that it is simply not a high priority for cadets in assessing the meaning of professionalism, a finding that may be more consistent with the conflicted Huntingtonian framework. Huntington contends that professionalism will naturally reinforce a commitment to an apolitical stance in domestic politics. Cadets' rank-ordering of technical expertise so high could mean the lessons they most internalize from Huntington relate to the autonomy accorded to the military to develop their skills, but in prioritizing this, they have not internalized the condition that Huntington says enables such autonomy and expertise in the first place-the military being apolitical. Huntington does not encourage reflection or discussion among military personnel of why remaining apolitical with respect to partisan politics is so important (Brooks 2020). This may be reflected in the results, which suggest a low prioritization of a value that is today, according to many military leaders, a defining feature of American military culture.
Notably, in his videotaped apology following his appearance in a photo-op with President Trump in Lafayette Square, Chairman of the Joint Chiefs, General Mark Milley reflected, "...we must hold dear the principle of an apolitical military that is so deeply rooted in the very essence of our Republic. And this is not easy. It takes time and work and effort. But it may be the most important thing each and every one of us does every single day." 23 The public debate over the politicization of the armed forces that followed the Lafayette Square photo-op-punctuated by the teaching moment of the Chairman's sincere apology-may underscore a broader theme of professionalism that is also evident in cadet responses: the officer corps may not truly reflect on the importance of being apolitical until it finds that norm in jeopardy.

List Experiment
In the survey's experimental portion, we utilized the "item-count" or list experiment techniques to probe respondent attitudes on controversial civil-military issues. This technique has been used to measure responses to issue areas in which "social desirability bias" may pose a problem to accurate measurement, such as racism or sexism (Sniderman and Carmines 1997, Gilens, Sniderman, and Kuklinski 1998, Corstange 2009 in the field breaking through similar desirability bias concerns (Kleykamp, Hipes, and MacLean 2017). As we expect certain areas of civil-military affairs will be controversial to future officers in the principal phase of their socialization to these organizational norms, we employ this technique to achieve the same result.
Rather than ask the entire sample a direct question to which the answer may be unreliable, we provide a list of statements, asking respondents to state with how many items do they agree. Our survey randomized respondents into both control and treatment groups. The control group receives a baseline list of 5 items, while the treatment group receives that same list, plus a controversial item around which the analysis is based. Since we believe that these future officers are likely to be highly sensitive to the notions of civilian control, unified command structures, and following orders, we expect that introducing the prospect of resisting these orders should be a subject about which they are unlikely to be honest. The experimental module was presented as follows (in randomized order), with the controversial item received only by the treatment group italicized by the authors: Now I am going to read you six statements about government, politics, and the military. After reading all six statements, tell us HOW MANY of them you agree with. For example, if you agree with three items on the list, enter choice "3". Remember, we will not know which answers you agree with, just how many.
1. I believe it is important for Congress to conduct oversight of military spending. 2. I believe the government should provide a universal health-care program for its citizens. 3. I believe the military should make the climate challenge a top priority for national security. 4. I believe it is acceptable for the President to use the military to enforce immigration policy. 5. I believe the country would benefit if more people were deferential to authority.

I believe it is important for the military to resist civilian orders that threaten the country's democratic traditions.
In order to provide a statistical baseline for subsequent analysis, after the control group provides an item-count response, we directly ask them the controversial item regarding resisting civilian orders. This should in many cases provide a 'lower-bound' for agreement with the controversial item, as we expect that respondents that are directly asked are more likely to conceal a willingness to push back on orders from civilian leaders than those that conceal those preference in the treatment group.
In a fashion similar to Kleykamp et al. (2017), Figure X displays the principal statistics and results of this experimental module. The nature of the list experiment prevents individual-level inference due to its design, which allows for item counts to conceal specific answer choices; as such, we use several established and newly-developed techniques to understand not only the extent of preference falsification, but demographic predictors for it.
As is customary for list experiments, we first conduct a simple difference in means calculation between the item counts from both groups, listed as "mean item counts". The difference between these figures ("true %"), is the first important baseline statistic for our consideration. This represents an estimate of the share of the population that would answer the controversial item honestly, if they were relieved of social desirability concerns. We then compare the true value to the direct value, the proportion of respondents who actually answered in the affirmative to the controversial item when asked. The difference between the two values represents the effect size of the experiment, the degree to which support for the controversial item is over-or under-expressed by the sample. In examining the results from Table 3, two significant findings emerge: first, the effect size of the experiment is high, at 32.9%, and second, it is in the opposite of the expected direction.
Despite our expectation that respondents would disagree with resisting anti-democratic civilian orders when asked directly, but express a willingness to do so when given the concealment of the treatment group, the results indicate the opposite. While 72% of respondents directly agreed with this idea, the true percentage value reveals that only 39% would actually agree with the statement.
The size of the effect and its unexpected direction are both remarkable findings in this experiment, leading us to investigate further.
Using newly-developed computational statistics techniques for analyzing item-count experiments, we explored what individual-level characteristics might predict this type of preference falsification. Using the list package in R, developed and employed in Imai (2011), we conduct a series of multivariate regression models designed to measure which of these predictors most strongly predict alignment with the controversial item. While a list experiment would typically look for strong positive coefficients in this regard, the direction of effects from our survey points us to coefficients with significant negative coefficients .  Across all model specifications, partisan identity and expressed political ideology display larger and more significant coefficients than considerations of race or gender.
Using zip code information provided by the respondent, we were able to geo-locate each individual's Congressional district (using the larger area for zip codes across multiple districts) and merge this information to district-level political data from the Cook Partisan Voting Index (CPVI). 24 As we noted above, given the direction and magnitude of effects from the experiment, regressors with high negative coefficients predict where preference falsification was likely to emerge. A binary indicator variable for the winner of the 2016 presidential election in the individual's home district proved a consistently negative and significant predictor against agreeing with the sensitive item.
In addition to the robustness provided by multiple model specification in  This finding is striking, in two ways. First, it suggests that large numbers of cadets consistently prioritize following orders over respect for the democratic institutions they will soon take an oath to protect. Second, it suggests that cadets realize that they are violating the spirit of that oath. When asked directly, cadets know the normatively appropriate response and respond accordingly. Yet, when provided an opportunity to reveal their true preferences through the mechanism of the list experiment, they indicate a clear lack of regard for that norm.
How can we make sense of this alarming finding? One possibility is that cadets simply care more about keeping their jobs, than for upholding the principles of democracy. Yet, the fact that there is a partisan skew in responses to the experiment cast doubt on this hypothesis. It is unclear why Republicans should care more about their jobs than other cadets. Something more seems to be afoot than self-interest.
A second possibility is that it has something to do with the ideological inclinations of cadets.
Conservative political philosophies are associated with more regard for hierarchy or authoritarian forms of social order. Recent survey research by Drutman, Goldman, and Diamond (2020) found that while few Americans support authoritarian alternatives to democracy, Republicans are more likely than Democrats to favor a strong leader in the abstract and unilateral action by the president.
Similarly, Federico, Feldman, and Weber (2017) found that Republicans scored 26 percentage points higher than Democrats on American National Election Studies' authoritarianism index in 2016.
A third possibility is that we are seeing more evidence of selective adherence to norms.
When able to keep their views private, cadets who identify as Republicans are especially likely to discount the importance of democratic norms, perhaps because they view them as antithetical to the arguments or positions held by their co-partisan president at the time the survey was taken. Given that Donald Trump's critics often decry his actions as contrary to democratic practices and traditions, cadets may privately reject adherence to those norms out of partisan support for the president. This is a significant finding because it suggests that partisan affinity is more important for many cadets than their professional commitments, even to the point of discounting the importance of a core tenet of their future oath. This finding is also significant in light of other findings that suggests selective adherence to other norms among some self-identified Republican cadets, such as their favorable response that one party is better at national security, and that "water cooler" talk about politics is okay in work contexts.
A final, not incompatible, possibility relates a core concern of the renowned sociologist, Morris Janowitz. 25 Janowitz argued that it was imperative that the military maintain the ethos of a citizen-solider. Doing so was essential to ensure it remained connected to society and supportive of the virtues of the republican form of government. He argued doing so was possible even for a professional military partly through the civilianization of the military and broadening the base of officer recruitment, but also through explicit training and education. Burk has argued that Janowitz's model was unrealistic; in the absence of mass mobilization of an army from the citizenry it would be difficult to sustain these core values. 26 Others have argued that the self-selection to military service that occurs with the all-volunteer and the demographic gaps between society and the U.S. military intensifies the social distance between the military and American society, manifesting in a civilmilitary gap. As William Rapp 27 and Robert Gates have articulated, the concern is that the military will become distant from society, and consequently, its personnel will come to view themselves as better or superior to average Americans. As Gates put it in a 2011 speech at West Point, "it is offputting to hear, albeit anecdotally, comments that suggest that military is to some degree separate and even superior from the society, the country, it is sworn to protect." 28 Janowitz reflects a deeper current of historical thinking: fears that professionalization could lead to a sense of distance. This is one reason why late nineteenth century advocates of a citizen- Hence, one way to interpret the results of the experiment is that Janowitz's worst fears have been realized. He was correct to worry that professionalism, especially of the Huntingtonian variety, would erode respect for democratic institutions and practices. Perhaps of all our findings, this looms as potentially the most unsettling. notion that military leaders should have equal footing in strategic decision-making and cadets' overwhelming endorsement of more retired GOFOs in the cabinet. There is more support for the conflicted Huntingonian model, in which respondents formally endorse Huntington's prescriptions, but then simultaneously express contradictory attitudes at odds with them with respect to civilian authority and engagement with politics. This is evident in the 36 percent who support politicians making partisan comments to military audiences, or the more than one quarter that think it is fine for retired officers to make partisan comments during elections. There is also significant evidence of selective adherence, in which partisan identity shapes when cadets support particular norms; cadets seem to engage in significant partisan rationalization when choosing to endorse particular norms.

Conclusions
Whether one focuses on the conflicted Huntingtonian model, or the selective adherence approach, however, it appears that socialization of West Point cadets is lacking with respect to their views about the military's engagement with partisan politics-a finding we surmise is not unique to West Point, but also probably evident in other pre-commissioning sources and the Army officer corps at large. In addition, the list experiment's stunning findings about the primacy of following civilian orders over maintaining the country's democratic traditions-and even more importantly, that cadets knowingly obscure this preference-suggests a superficial appreciation, if not blatant disregard, for the Constitution to which cadets will upon their commissioning take an oath to uphold.