PLACEHOLDER NOT UPDATED!!!

Introduction

Self-control, self-discipline, willpower… these words have practically taken up residence in everyone’s New Year’s resolutions and daily anxieties. But are they all the same thing?

This is, in fact, a series of complex concepts. If we want to discuss this topic clearly and productively, the prerequisite is to first clarify these ideas and terms, to understand what we are actually discussing.

This article, drawing on a 2021 psychology review paper (Inzlicht et al., 2021), attempts to summarize what we commonly refer to as “self-control.”

As a side note, I will also try to create an (informal) correspondence between some of the answers under [the Zhihu question “How to improve self-control?”](_link placeholder_) and the theoretical models from the literature.

Finally, I’ll share some of my personal speculations: Why do people care so much about self-control? And why am I personally grappling with this issue?

Alright, without further ado, let’s get started!

Main Text

First, a huge and formal thank you to: Integrating Models of Self-Regulation (Inzlicht et al., 2021). Truly the work of leading experts, organizing the tangled mess of various theories with such clarity. I felt so refreshed after reading it!

What Does This Paper Say?

Clarification: Self-Regulation vs. Self-Control vs. Cognitive Control

Before analyzing the theoretical models, we must first clarify three core terms that are often used interchangeably.

  • Self-Regulation

    This is a broad “Umbrella Term.” It refers to the entire dynamic process by which an individual sets goals, guides their thoughts, emotions, and behaviors toward those goals, and monitors and adjusts their progress. It covers all stages from goal setting to execution, monitoring, and adjustment—it’s the “full-cycle management” of achieving a goal.

  • Self-Control

    This is a subset of self-regulation, closer to what we often call “willpower.” It specifically refers to the process of resolving conflicts and inhibiting impulses when temptations and goals compete. Self-control is only activated when conflict and temptation arise, used to overcome immediate desires in service of long-term goals.

  • Cognitive Control

    This is primarily a concept from cognitive neuroscience, referring to the underlying cognitive abilities that serve goal-oriented behavior. It includes flexibly allocating attention, inhibiting irrelevant information, and updating working memory. It is closer to the brain’s underlying “executive function” or “information processing efficiency” and is not entirely equivalent to “willpower” (self-control) in everyday situations.

Seven Theoretical Models: How Do We Self-Regulate?

Scholars have proposed various models from different perspectives to explain how individuals regulate themselves and why “self-control failure” occurs.

1. Cybernetic Control Model

This model views self-regulation as a feedback loop, much like a thermostat. The individual constantly compares the “goal” and the “current state.” Once a discrepancy is detected, action is automatically taken to reduce this gap.

2. Goal Systems Theory

This model focuses on the “network of relationships” between goals and the means to achieve them. The relationship between means and goals affects motivation. For example, when you find a means that “kills two birds with one stone” (Multifinal Means) (e.g., “cooking for a_self” is both “saving money” and “being healthy”), your execution efficiency may be highest.

3. Dual Systems Models

This model posits that self-control is the result of a “tug-of-war” between two internal systems: one is the “emotional, impulsive” System I (the hot system), which seeks immediate gratification; the other is the “rational, deliberate” System II (the cold system), which focuses on long-term goals. Self-control failure is seen as System I overpowering System II in certain contexts (like fatigue or stress).

4. Choice Models

Core Idea: This model argues that self-control is not a “tug-of-war” between two systems, but rather a (seemingly) rational “value calculation.” The individual compares the “subjective value” (i.e., expected benefit minus expected cost) of all options and always chooses the one with the highest subjective value at that moment.

Real-life Example: Take the choice between “going to bed early or scrolling short videos” at night.

  • Option A (Scrolling short videos):

    • Benefit: Immediate pleasure (+100)

    • Cost: Tired tomorrow (-50), Anxiety from delaying long-term goals (-30)

    • Subjective Total Value: +20

  • Option B (Putting down the phone and sleeping early):

    • Benefit: Energetic tomorrow (+40), Pride in achieving a goal (+50)

    • Cost: Boredom right now (-20), Effort to resist temptation (-10)

    • Subjective Total Value: +60

Decision and Failure: In this calculation, because +60 > +20, the individual chooses to “go to bed early.”

So how does self-control failure happen? Because this “value calculation” is subjective and unstable. For example, if the individual is extremely tired or stressed today, the benefit of “immediate pleasure” might skyrocket to +200, while the benefit of “energetic tomorrow” devalues to +5. At this point, the calculation reverses (+170 vs +25), and the individual will choose to scroll videos.

Therefore, people with so-called “strong self-control” may simply be those who, in their subjective value calculations, habitually assign a higher weight and value to “long-term benefits.”

5. Resource Model / Ego Depletion

This model compares willpower to a finite “battery” or “muscle” resource. Each time self-control is used (like enduring stress or resisting temptation), the resource is depleted (i.e., “Ego Depletion”). When the resource is exhausted, the individual is more likely to fail at self-control in subsequent tasks.

(Note: Although this model is extremely well-known, it is currently highly controversial in academia due to its low replicability.)

6. Process Model of Self-Control

Core Idea: This model suggests that the key to successful self-control is not “brute-forcing” it (i.e., strength of willpower), but rather the use of strategies. The timing of the strategy intervention is decisive: the earlier the intervention (prevention), the less willpower is consumed and the better the effect; the later the intervention (responding to the temptation), the more it relies on “brute-force” willpower, and the more likely it is to fail.

Real-life Example: Suppose a person’s goal is to eat healthily, and their biggest temptation is the high-calorie snacks colleagues bring to the office. They can use strategies at the following 5 stages, in descending order of effectiveness:

  1. Strategy 1: Situation Selection (Most Effective)

    • Action (Prevention): Before the colleague brings out the snacks, choose to work in a snack-free conference room, or apply to work from home.

    • Effect: Avoids the temptation at its source. No willpower is consumed.

  2. Strategy 2: Situation Modification

    • Action (Prevention): Since leaving isn’t an option, when the snacks appear, proactively close the snack box lid, or move it to a corner where it can’t be seen.

    • Effect: The temptation still exists, but accessing it is harder, consuming only a small amount of willpower.

  3. Strategy 3: Attentional Deployment

    • Action (Intervention): The snacks are already on the table. The individual forces themselves to shift their attention, for example, by immediately turning to their computer screen or starting a work-related discussion.

    • Effect: Requires willpower to actively shift attention and avoid “seeing” the temptation.

  4. Strategy 4: Cognitive Change

    • Action (Intervention): The individual sees the snack and mentally “re-appraises” it. They tell themselves: “This isn’t delicious food; it’s trans fats and artificial sweeteners, it’s artery-clogging garbage.”

    • Effect: The temptation’s allure is subjectively reduced, but this requires significant cognitive effort to “convince” oneself.

  5. Strategy 5: Response Modulation (Most Ineffective / Most Tiring)

    • Action (Intervention): This is the “head-on confrontation” of willpower. The individual has already picked up a cookie, brought it to their mouth, and at the very last second, uses “immense willpower” to force themselves to put it back down.

    • Effect: Consumes a huge amount of willpower resources, and the failure rate is extremely high.

This model emphasizes that truly disciplined people are not those with the strongest willpower (i.e., best at Strategy 5), but those who are best at using Strategy 1 and Strategy 2 to avoid putting themselves in situations that require willpower in the first place.

7. Trait Models of Impulsivity

This model doesn’t focus on the “process” but on “types of people.” It argues that “impulsivity” (the opposite of self-control) is not a single trait but is composed of different personality traits (like the UPPS model). For example, some people are impulsive because they struggle to control themselves during negative emotions (high Urgency), while others are impulsive simply because they lack forethought (low Premeditation).

Theory and Practice: Which Models Do the Zhihu Answers Correspond To?

Below, I’ve selected some answers from [the original Zhihu post](_link placeholder_), extracted some of their points, and tried to map them to the theoretical models above. This is an (informal) look at the different angles from which people are approaching this topic:

Extracted Point 1: “Two Curves” & Value Integration Formula

“The further in the future something is, the weaker its present value and motivation… Under this short-sighted weight distribution, the short-term advantage of scrolling on a phone… the integrated result will, in fact, be far higher than studying.”

Model Analysis:

  • Primary Correspondence: Choice Models

    This perfectly illustrates the single-valuation process of the “Choice Model.” The “short-sighted weight distribution” mentioned is known in economics as “hyperbolic discounting,” which is the core mechanism this model uses to explain self-control failure: we place a much higher value on immediate rewards (scrolling) than on delayed rewards (studying).

  • Secondary Correspondence: Dual Systems Models

    This “short-sighted” valuation system can also be understood as “System I” (the impulsive, hot system) dominating the decision, while the “rational” integration (System II) is suppressed.

Extracted Point 2: “National Focus Tree” & “Recursive Stable-State Iteration Protocol” (RSIP)

“Must wash the dishes as soon as possible after eating at home… to avoid falling into a larger state of lethargy.” And “The closer to the root, the more child nodes a routine has, the greater the cost of losing it… slowly internalizing it into your habit.”

Model Analysis:

  • Correspondence: Process Model of Self-Control

    “Washing the dishes as soon as possible” is a preventive strategy, specifically, a form of “situation modification.” By changing the environment (getting rid of the dirty dishes), it removes a subsequent temptation (e.g., “the kitchen is such a mess, I’ll just skip studying”). This aligns with the core idea of the “Process Model,” which is that self-control relies on strategy, not brute force.

(Note: This is just a simple correspondence. The meaning and technical highlights of the “Recursive Stable-State Iteration Protocol” & “National Focus Tree” will be analyzed further in the next article.)

Extracted Point:

“If a task is not yet completed… the desire for cognitive closure will motivate us to return to the unfinished task.”

Model Analysis:

  • Correspondence: Cybernetic Control Model

    This is a textbook definition of the “Cybernetic Control Model.” The core of this model is a feedback loop. “A task is not yet completed” means the monitoring system has detected a “discrepancy” between the “current state” and the “goal state.” The “desire for cognitive closure” is the motivation that drives the “operating system” to act to reduce that discrepancy.

Extracted Point:

“Every act of self-control diminishes our focus on long-term, disciplined goals, gradually leading us to make a series of hasty and rash decisions.”

Model Analysis:

  • Correspondence: Resource Model / Ego Depletion

    This is a classic description of the “Ego Depletion” effect. This model posits that self-control relies on a limited resource. “Every act of self-control” “diminishes” this resource, leading to “diminished focus on long-term goals” and “hasty decisions” (i.e., a decline in self-control performance) in subsequent tasks.

Extracted Point:

“Recording means visualizing the results. The human brain, upon seeing the results, will automatically optimize the method and path.”

Model Analysis:

  • Correspondence: Cybernetic Control Model

    “Recording” and “visualization” strengthen the key steps in the “Cybernetic Control Model”:

    1. Input: “Recording” provides the brain with clear input about the “current state.”

    2. Monitor: The brain “seeing the results” is the act of comparing the “current state” with the “goal state.”

    3. Implement: The brain “automatically optimizing the method and path” is the “operating system” taking action to reduce the discrepancy.

Extracted Point:

“In reality, there aren’t that many things in the world that require ‘control’ to do… Fixating on ‘self-control ability’ is the wrong approach, and it paradoxically makes you lose your self-control.”

Model Analysis:

  • Correspondence: Overemphasis on “Conflict” (A reflection from the review paper)

    This viewpoint critiques the mainstream self-control theories’ excessive focus on “conflict.” This aligns with the review authors’ own reflections that the past focus on “resolving conflict” might be detrimental, and more attention should be paid to “conflict-free self-regulation.” When people pursue goals for “autonomous reasons” (i.e., they are “fully focused on doing it well”), they experience fewer temptations and conflicts, and thus succeed without needing “self-control.”

Extracted Point:

“The experimenter’s mindset—observe, record, allow for mistakes, use ‘your original self’ as the reference.”

Model Analysis:

  • Correspondence: Cybernetic Control Model

    This also describes the “Cybernetic Control Model’s” feedback loop. “Observe” and “record” are “Monitoring.” “Using ‘your original self’ as the reference” is setting the “Current state.” And “allowing for mistakes” is the correct understanding of this loop: a “mistake” is just a “discrepancy” detected by the monitoring system. It’s not an endpoint, but a signal that triggers the “operating system” to make an adjustment.

[My Own Thoughts]

Extracted Point:

“Trying to find the key decision points through continuous ‘self-awareness’… I found this path had more drawbacks than benefits: not only did it occupy valuable cognitive resources, it also exacerbated anxiety…”

Model Analysis:

  • Correspondence: Cost of Cybernetic Control / Role of Emotion

    This is a reflection on the cost of the “Monitoring” step in the “Cybernetic Control Model.” The literature mentions that conflict itself is aversive and causes anxiety. This kind of high-intensity “meta-monitoring” (“continuous self-awareness” of one’s own decision-making process) is, in itself, a high-cost cognitive activity. It “occupies valuable cognitive resources” and “exacerbates anxiety.”

Why Cares? — Why Do We Care About “Self-Control”?

What: What Are People Discussing? (Three Perspectives)

When people discuss “self-control” and related topics, they are actually looking for answers at three different levels of analysis. The picture becomes very clear when we map these perspectives to the theoretical models:

1. Macro Perspective: “Am I just an undisciplined person?” (Trait Theory)

This perspective views self-control as a stable personality trait. The discussion tends toward “attribution,” such as “I was just born an undisciplined person.”

  • Corresponding Model: Trait Models of Impulsivity.

  • Analysis: This level focuses on an individual’s personality dimensions. For example, analyzing whether a person “always stays up late on their phone” is because they are prone to impulsivity under negative emotions (high Urgency) or simply because they lack advance planning (low Premeditation).

2. Meso Perspective: “How should I design a disciplined life?” (Strategy Theory)

This perspective treats self-control as a systems engineering or project management problem. It focuses on “methodology,” “habit-building systems,” and “environmental design.”

  • Corresponding Models: Process Model of Self-Control & Goal Systems Theory.

  • Analysis: This is the “discipline architect” perspective. On one hand, people discuss “uninstalling the game” or “charging the phone in the living room,” which correspond to “Situation Selection” and “Situation Modification” strategies in the “Process Model”—that is, avoiding conflict at the source. On the other hand, people look for “multitasking” actions, like “cooking for oneself to be healthy and save money,” which corresponds to finding efficient “multifinal means” in “Goal Systems Theory.”

3. Micro Perspective: “Why did I lose control ‘in this moment’?” (Battle Theory)

This perspective views self-control as an “immediate battle” with temptation. It focuses on “Why did I pick up my phone again?” or “Why couldn’t I refuse that piece of cake?”

  • Corresponding Models: Dual Systems Models, Choice Models, Resource Model, Cybernetic Control Model.

  • Analysis: This level attempts to explain the “in-the-moment” failure:

    • “My rational and emotional sides are fighting”—This is the Dual Systems Model (System I vs. System II).

    • “I calculated it, and the pleasure of scrolling is worth more right now”—This is the Choice Model (the value of immediate rewards is overestimated).

    • “I’m too tired from a long day at work, I just don’t have the energy to go to the gym”—This is the Resource Model (willpower battery is depleted).

    • “I need an app to track my progress”—This is the Cybernetic Control Model (attempting to optimize the feedback loop through recording and monitoring).

Why: Why Do People Care? (Four Motivations)

Behind people’s obsession with “self-control” are several different core value pursuits.

1. Pursuit of Excellence and Performance (Performance Motive)

From this perspective, self-control is the “fuel for success,” seen as a productivity tool to maximize personal output in order to win in academic, professional, or financial competition (including “involution”). This motive aligns with the “work performance,” “financial savings,” and other outcomes of discipline mentioned in the literature.

2. Alleviating Pain and Anxiety (Pain-Relief Motive)

The starting point for this motive is not “to become stronger,” but “to stop the pain.” For example, severe procrastination has already led to tangible negative consequences (like poor academic performance) and a strong psychological burden (like guilt and self-loathing). People seek self-control to escape this negative emotional state caused by conflict and failure.

3. Seeking Meaning and Identity (Identity Motive)

This motive is more philosophical. Self-control is seen as the bridge to “achieving one’s ideal self.” People care about “Am I living out the values I cherish?” For example, someone persists in working out, not for competition or to relieve anxiety, but because “health” is a core value. Self-control failure, in this case, is seen as a “betrayal of the self.”

4. Desire to Understand and Control (Control Motive)

This is a “hacker mindset” or “experimenter mindset,” fascinated by the ultimate proposition of “understanding the self,” attempting to “model” and “systematize” one’s own brain and behavior. This motive is highly consistent with systems thinking, like the “Cybernetic Control Model,” trying to transform a chaotic psychological process into a predictable, optimizable engineering problem through “observation,” “recording,” and “iteration.”

Where to Go? — A Small (and Arrogant) Prediction

As a (humble) master’s student interested in HCI x Well-being, I’m going to make a bold prediction about the future!

With the progress of our times and the development of AI, more and more people are paying attention to fields like Personal Informatics, Self-Regulation, mHealth, and Positive Computing. Essentially, they hope to use information technology to promote individual well-being (of course, used in the wrong way, it could also exacerbate involution~).

If we limit the problem to “self-regulation” (a term I personally prefer, as it’s more neutral and not tied to “self-discipline” or “self-control,” which often imply anti-human “self-torture” or “involution”), I believe:

  1. First, more and more people will notice the issue of value-driven pursuits. After all, a self-regulation app can be used to promote “involution,” but it can also be used to promote “creativity” and “reduce suffering.”

  2. Second, self-control will gradually move toward externally-assisted “other-control.” With the development of fields like personal informatics and HCI, and the emergence of various tools, people will need to spend less effort on self-monitoring. At the same time, AI, just as it does in programming and writing, can help individuals reduce the cognitive cost of decision-making and provide more effective interventions.

    I hope I can be fortunate enough to contribute a little something to this trend.

Concluding Rambles

Two years ago, I was inspired by an article from [@叶峻峣](_link placeholder_), and the idea of “porting” and organizing articles related to psychology was born. However, time flew by, and my output has been 0… (facepalm)

A month ago, I saw [@Edmond’s](_link placeholder_) article on self-control and was once again given a “kick in the pants,” feeling that I couldn’t procrastinate any longer! So, I (finally) started writing!

Next up, I plan to discuss and analyze some specific technologies, including commercial apps and academic papers I’ve encountered. I won’t procrastinate (gū gū) this time!