Working Through Moral Psychology at Scale
I spent more time than I want to admit trying to map Jonathan Haidt's framework onto actual organizational behavior studies, and Chapter 10 turned out to be the one that broke my model a little. The chapter, The Righteous Mind Chapter 10 Summary territory, deals with the Google case study and how moral communities shift when incentives change. It is less about Google specifically and more about what happens when a group's dominant moral foundation gets crowded out by something else. Haidt uses Google as a live example of moral drift in a high-status institution. The company started with "don't be evil" as an informal moral anchor, which aligned pretty cleanly with the caring and fairness foundations that tend to dominate liberal moral psychology. Over time, the pressure to monetize and scale pushed the organization toward a different configuration. Not evil, exactly. Just... optimized differently. Haidt points out that this is not a failure of character. It is a structural feature of how group-level morality works. The core argument rests on Haidt's Moral Foundations Theory. The six foundations are care/harm, fairness/cheating, loyalty/betrayal, authority/subversion, sanctity/degradation, and liberty/oppression. Different political and institutional groups weight these differently. When you are watching a company change its moral vocabulary, you are watching its foundation profile shift. Google moved from a care-and-fairness-dominant posture to something that leaned harder on liberty and, eventually, on raw utility calculations that do not map cleanly onto any single foundation.
What most readers miss is that Haidt is not making a normative claim here. He is not saying Google got worse. He is showing that moral systems are adaptive, and adaptation can look like betrayal to people who were onboarded under the old configuration. That distinction matters because it changes how you analyze any institution. I ran into a specific problem when I tried to apply this framework to a mid-size tech firm that had gone through a similar transition. We had early internal memos, product ethics guidelines, and public statements from the first five years. The later years told a completely different story, but the language was almost identical. Words like "trust," "community," and "doing the right thing" appeared in both eras. The foundation profile had shifted underneath the vocabulary, which made any straightforward content analysis useless. I ended up having to code the documents against each foundation individually, track the frequency of foundation-adjacent terms across years, and then cross-reference that with actual policy changes and executive compensation structures. The gap between stated values and operational incentives is where the real moral shift happens, and you cannot see it from press releases alone. The workaround was to treat the moral language as signal noise and focus on incentive structures instead. Who got promoted when the language changed. Which metrics replaced which ones. Where the budget moved. That gives you a much cleaner picture of what actually changed than any rhetorical analysis will.
How the Chapter Fits Into the Larger Argument
Chapter 10 sits in the book's final section, which is where Haidt moves from describing moral psychology to proposing how we might actually talk across moral divides. The Google case is a bridge. It shows that moral communities are not static, that they respond to environmental pressures, and that people inside the community often cannot see the shift happening because they are immersed in it. That is why the surrounding chapters on moral mysticism and group-level selection matter. Individual-level moral reasoning does not explain institutional drift. Group-level dynamics do. Haidt also uses this chapter to set up his later point about righteousness. When a group's moral foundation changes, the people who anchored themselves to the old configuration do not just disagree with the new direction. They experience it as moral corruption. That is not bias. It is how moral intuition works. Your gut response to a values shift is immediate and visceral, and it predates any rational analysis you could run on it. This creates a practical problem for anyone trying to analyze organizational ethics. You have to separate the intuitive response from the structural one. Most people conflate them. They see a company change its stance and immediately label it as moral decline without checking whether the underlying incentive structure actually changed or whether the surface language just failed to keep up.
Get the Full Details
Common Pitfalls When Applying This Framework
The biggest mistake I see is treating the six foundations as political checkboxes. Care equals liberal. Loyalty and sanctity equal conservative. Liberty equals libertarian. That mapping is too rough for anything beyond casual conversation. Haidt himself warns against this simplification, but it keeps coming up in summaries and secondary analyses because it is convenient. Another pitfall is assuming that foundation shifts are conscious decisions. They almost never are. They are emergent properties of compensation structures, hiring patterns, investor pressure, and competitive dynamics. The language changes first because language is cheap. The behavior changes later because behavior costs money. If you are tracking a moral shift, start with the money. A third issue is the temptation to find symmetry where none exists. Liberal institutions do not drift the same way conservative ones do, and vice versa. The directions of drift are different because the starting foundation profiles are different. Google's drift toward utilitarian optimization is not the same type of shift as, say, a religious organization drifting toward secular compliance. The foundations involved overlap, but the mechanics are distinct. Treating them as equivalent muddies the analysis.
What the Chapter Gets Wrong or Leaves Out
Haidt's Google case study is illustrative, not exhaustive. He does not have access to internal meetings, compensation committees, or board-level debates. The analysis is built from public records and observable outcomes, which means it can miss the actual decision points where the shift became irreversible. In my experience, those moments are usually unglamorous. A budget reallocation. A key hire with different incentives. A metric change that quietly redefined success. He also does not fully address the role of external criticism in accelerating moral drift. When an institution faces sustained moral scrutiny from outside, it sometimes responds by doubling down on its original foundations and sometimes by shifting toward whichever foundation is least likely to attract criticism. The Google case shows the latter pattern, but Haidt does not spend much time on the feedback loop between external pressure and internal moral reconfiguration. That is a gap worth noting if you are applying this to current events. Another limitation is the chapter's reliance on a single case study for a broader claim. One company's trajectory does not establish a general law. You need multiple cases across industries, sizes, and political contexts to know whether the pattern holds. Haidt acknowledges this implicitly by positioning the chapter as an illustration rather than a proof, but readers sometimes miss that distinction.
Practical Takeaways
If you are using this framework to analyze an organization, start by mapping its current foundation profile against its founding documents. The gap between the two tells you more than either snapshot alone. Then track incentive structures, not language. Compensation, promotion criteria, and metric definitions reveal the actual moral priorities. Finally, resist the urge to call a shift "corruption" or "progress" before you have checked whether the underlying group-level dynamics support that judgment. The Google case works precisely because Haidt withholds that judgment and lets the structural analysis speak for itself. The chapter is worth reading not for its diagnosis of Google but for the method it models. Moral analysis at the group level requires looking past rhetoric to incentive structures, recognizing that shifts are often invisible to participants, and accepting that different moral foundations are adaptations, not errors. That is a harder habit to maintain than taking a side, but it is the only one that produces useful results.
