5.7 Modeling Assumptions and Limits of Common Force Laws

When the Formula Fails

Your textbook says friction is $f = \mu N$. This is a clean, satisfying formula. Two quantities multiplied together, one dimensionless constant, and you have the friction force. It works beautifully for a wooden block on a wooden ramp, for a crate sliding across a warehouse floor, for homework problems involving inclined planes.

But try it on a tire gripping wet ice. Try it on a gecko climbing a glass wall. Try it on a rubber eraser pressed against a polished tabletop, where friction actually increases with contact area --- something the formula says should not matter at all. Try it on a curling stone gliding across ice, where the friction force depends on how long the stone has been sliding.

The formula does not just become slightly inaccurate in these cases. It gives qualitatively wrong predictions. It says friction should not depend on area --- but sometimes it does. It says kinetic friction should not depend on speed --- but sometimes it does. It says the friction coefficient is a property of the two surfaces --- but sometimes it depends on temperature, humidity, surface contamination, and the history of contact.

Does this mean the formula is wrong? Not exactly. It means the formula is a model --- an approximation that captures the dominant behavior within a specific domain of conditions. Outside that domain, the model breaks down, and the breakdown itself is informative. It tells you something new about the physics.

This section is about developing that meta-skill: not just using force models, but knowing when to trust them and when to question them.

Prediction

Before you read on: The kinetic friction model $f_k = \mu_k N$ says that friction does not depend on the speed of sliding. Consider three scenarios:

(a) A car's tires sliding on asphalt at 200 km/h

(b) A curling stone gliding across ice at 3 m/s

(c) A ski sliding over packed snow at 20 m/s

For each one, do you think the constant-friction model is reliable? Where would you most expect it to fail, and why?

Commit to your answers. There is no single right response here --- the goal is to engage your physical intuition before we examine the evidence.

The Guiding Question

When do the standard textbook force models begin to fail --- and what does the failure tell us?

This is a different kind of question than the ones you have been answering in this chapter. Sections 5.1 through 5.6 asked you to use force models: draw the FBD, write $\sum \vec{F} = m\vec{a}$, solve. This section asks you to step back and examine the models themselves. Every force law you learned in Section 5.4 --- weight, tension, normal force, friction, drag --- is an approximation. The question is not whether it is approximate (it always is), but where the approximation breaks.

Exploration: Watching a Model Break

Consider the simplest friction experiment: a block sliding across a surface at varying speeds.

[Interactive: Friction vs. Speed. A block slides across a surface. The student controls the initial speed using a slider (from 0.1 m/s to 50 m/s). Two curves are displayed on a graph: the model's predicted friction force ($f_k = \mu_k N$, a horizontal line) and the "measured" friction force (from empirical data). At low speeds (below about 1 m/s), the two curves nearly overlap. As speed increases, they diverge. The student can also switch between surface types --- wood-on-wood, rubber-on-concrete, metal-on-ice --- and see different divergence patterns. A guided prompt reads: "At what speed does the predicted friction start to deviate noticeably from the measured value? Does the measured friction increase or decrease with speed? Does the answer depend on the surface?"]

Spend time with this. Try all the surfaces. Pay attention to when the divergence begins and which direction it goes. For some surfaces, friction decreases at high speed (think of tires losing grip). For others, it increases (think of viscous effects at a lubricated interface). The constant-friction model misses all of this because it was never designed to capture speed dependence.

Pause and think: The model $f_k = \mu_k N$ has no variable for speed. Could you have predicted, just from looking at the formula, that it would fail at high speeds? What does the absence of a variable tell you about a model's assumptions?

That last question is key. When a model omits a variable, it is making an implicit claim: this quantity does not matter. The absence of speed in $f_k = \mu_k N$ is not an oversight --- it is an assumption. The model assumes friction is speed-independent. When that assumption fails, so does the model.

Concept Reveal: Every Model Has a Domain of Validity

Here is the central idea of this section, stated plainly:

Every force model is an approximation with a domain of validity. The domain is defined by the assumptions behind the model. When those assumptions are violated, the model breaks down --- and the nature of the breakdown points toward the physics that the model was ignoring.

This is not a weakness of physics. It is how physics works. Models are not supposed to be universally true. They are supposed to be useful within specified conditions. The skill is knowing those conditions.

Let's make this concrete by auditing the assumptions behind each force model from Section 5.4.

Weight: $W = mg$

This model says the gravitational force on an object near Earth's surface is constant --- it does not depend on altitude, location, or anything else.

Assumptions: - The object is near Earth's surface (altitude $\ll$ Earth's radius) - "Near" means $g$ does not vary appreciably over the relevant distance - The object is small enough that $g$ is uniform across it

Where it fails: - At high altitudes. A satellite at 400 km altitude experiences $g \approx 8.7$ m/s$^2$, not $9.8$ m/s$^2$. The "constant $g$" model is off by about 11%. - Over very large objects. For a suspension bridge spanning kilometers, $g$ varies slightly from one end to the other --- though this is almost never important in practice. - On other planets. This one is obvious, but students sometimes apply $g = 9.8$ m/s$^2$ reflexively even in problems set on the Moon.

What replaces it: Newton's law of universal gravitation, $F = Gm_1 m_2 / r^2$, which gives the full distance dependence. The constant-$g$ model is the first-order approximation of this law near the surface.

Tension in a String

The standard model says tension is the same throughout a massless, inextensible string.

Assumptions: - The string has negligible mass compared to the objects it connects - The string does not stretch - The string is not accelerating so rapidly that its own inertia matters

Where it fails: - Heavy ropes. If a thick climbing rope hangs under its own weight, the tension at the top is greater than at the bottom by the weight of the rope below. The "uniform tension" model fails completely. - Elastic cords. A bungee cord stretches, and the tension depends on the amount of stretch (Hooke's law). Treating it as inextensible gives the wrong physics. - Whipping motions. When you crack a whip, the tip accelerates to supersonic speeds. The string's mass and elasticity dominate the dynamics.

Normal Force

The model says the normal force adjusts to prevent interpenetration of surfaces, and its value is determined by Newton's second law applied to the object.

Assumptions: - The surface is rigid (it does not deform under load) - Contact is maintained (the object stays on the surface)

Where it fails: - Soft surfaces. When you step on a mattress, the surface deforms significantly. The normal force depends on the deformation, and the simple "rigid surface" picture breaks down. - Loss of contact. On a roller coaster at the top of a hill, if the car moves fast enough, the normal force drops to zero --- and the model predicts a negative normal force, which means contact has been lost. The model signals its own failure.

Kinetic Friction: $f_k = \mu_k N$

Assumptions: - Friction is independent of contact area - Friction is independent of sliding speed - The coefficient $\mu_k$ is a constant property of the two surfaces - The surfaces are dry (no lubricants)

Where it fails: - Speed dependence. At very high speeds, friction coefficients for many materials decrease (tires lose grip) or increase (viscous heating effects). At very low speeds, "stick-slip" behavior appears --- the surfaces alternate between sticking and sliding. - Area dependence. For soft, deformable materials like rubber, friction does depend on contact area because more surface deforms and interlocks. - Temperature and wear. Braking generates heat, which changes the surface properties and the friction coefficient during the process. - Gecko feet and adhesion. Gecko adhesion involves van der Waals forces at the molecular level. The friction model $f = \mu N$ is not even the right framework --- the force does not depend on a normal load.

Drag: $f_D = bv$ or $f_D = cv^2$

Assumptions (linear drag $bv$): - The object moves slowly enough that the flow around it is smooth (laminar) - The fluid is uniform and unbounded

Assumptions (quadratic drag $cv^2$): - The object moves fast enough that the flow is turbulent - The drag coefficient is approximately constant

Where it fails: - The transition region. Between laminar and turbulent flow, neither model is accurate. The drag force has a complicated dependence on speed. - Near boundaries. An object falling near a wall or in a confined tube experiences different drag than in open air. - Compressibility effects. At speeds approaching the speed of sound, the air compresses and the drag coefficient changes dramatically.

Connection: Models as Local Approximations

In Section 4.4, you studied linearization --- the idea that a complicated function can be approximated by a simpler one near a chosen point. The approximation is excellent close to that point and degrades as you move away. The key question was always: how far can you go before the approximation becomes unacceptable?

Force models work the same way. The model $f_k = \mu_k N$ is a kind of linearization of the true, complicated friction phenomenon. It captures the dominant behavior in a region of parameter space (moderate speeds, hard surfaces, dry contact) and ignores everything else. Move outside that region, and the model degrades --- just as a linear approximation degrades away from its expansion point.

This is also a callback to Chapter 1, where you first encountered the idea that physics proceeds by idealization. Every problem begins with choices about what to include and what to ignore. Now you can see those choices with more precision: each simplification corresponds to a specific assumption, and each assumption defines a boundary beyond which the model should not be trusted.

Metacognition: What Experts Know That Beginners Don't

Here is a striking fact about expert physicists: they spend more time thinking about whether a model applies than about how to solve the equations once it does.

A beginner sees a friction problem and writes $f = \mu N$ without hesitation. An expert asks: Is friction the right framework? Is the constant-coefficient model appropriate? Is the normal force large enough for the model to dominate over adhesion effects? Is speed low enough for the speed-independent assumption to hold?

These are not questions about mathematical technique. They are questions about the relationship between the model and the physical world. They require a kind of knowledge that is not captured in the formula itself --- meta-knowledge about the formula's origins, assumptions, and limits.

This is what separates textbook physics from real physics. Textbook physics gives you the formula and asks you to apply it. Real physics asks: does this formula even apply here? And if not, what does?

You are building that meta-knowledge right now, by examining the assumptions behind each model. It will not show up on a plug-and-chug problem. But it will show up every time you encounter a situation where the textbook model gives a prediction that conflicts with your experience or your data.

Why Simplified Models Are Still Valuable

After all this discussion of failure modes, you might wonder: why bother with these models at all?

Because they work astonishingly well within their domain. The constant-$g$ model is accurate to better than 0.3% for any altitude below 10 km. The friction model $f_k = \mu_k N$ predicts the behavior of sliding blocks to within a few percent for a huge range of everyday situations. The quadratic drag model captures the terminal velocity of a skydiver to better than 10%.

Simplified models are not "wrong" in any useful sense of the word. They are limited. And their limitations are well-characterized, which means you can estimate the error you are making by using them. That is a feature, not a bug. A model whose limitations you understand is far more useful than a complex model you cannot evaluate.

There is also a pedagogical reason. You cannot understand the corrections until you understand the baseline. The speed-dependent friction model $f_k = \mu_k(v) N$ only makes sense if you first understand the speed-independent version and then ask: what changes when speed matters? Every refined model is a departure from a simpler model. The simple model provides the reference point.

Spaced Retrieval

Before moving to practice, test your recall of earlier material.

Recall prompt 1: What is a free-body diagram, and why is it the essential first step before writing $\sum \vec{F} = m\vec{a}$? (Section 5.1)

Recall prompt 2: Newton's third law says forces come in equal-and-opposite pairs. If the Earth pulls you down with force $mg$, what is the reaction force, and what object does it act on? (Section 5.2)

Recall prompt 3: In Section 4.4, you learned about linearization as a way to simplify differential equations near a point. In one sentence, how is that idea related to what this section says about force models?

Practice Layers

Layer 1: Concrete -- List the Assumptions

Problem 1. You are analyzing a problem in which a 2 kg block slides down a 30-degree incline at moderate speed. You model the friction as $f_k = \mu_k N$ with $\mu_k = 0.3$. List three assumptions you are making about friction in this setup. For each assumption, briefly describe a scenario in which it would fail.

Check your answer

Three key assumptions (others are possible): 1. **Friction is independent of speed.** This would fail if the block were sliding at very high speed (e.g., a re-entry vehicle on a heat shield), where frictional heating changes the surface properties and the effective coefficient. 2. **Friction is independent of contact area.** This would fail if the block were made of soft rubber, which deforms under load and increases the true contact area, thereby increasing friction. 3. **The coefficient $\mu_k$ is constant throughout the motion.** This would fail if the surface were wearing down during the slide (e.g., sandpaper), changing the roughness and thus the friction coefficient as the motion progresses. The point is not that these failures happen in typical textbook problems --- they usually do not. The point is that you should *know* what you are assuming, so you can recognize when a real situation falls outside the model's domain.

Layer 2: Pattern -- Identify the Failing Assumption

Problem 2. For each scenario below, identify which assumption of the standard force model is most likely to fail.

(a) A skydiver reaches terminal velocity and then opens a parachute. You model drag as $f_D = cv^2$ with a single constant $c$.

(b) You model the tension in a heavy anchor chain as uniform throughout.

(d) You assume kinetic friction $f_k = \mu_k N$ for a car braking on a wet road at 150 km/h.

Check your answer

**(a)** The assumption that the drag coefficient $c$ is constant fails. When the parachute opens, the cross-sectional area and drag coefficient change dramatically. The value of $c$ before and after deployment are completely different. You would need to switch to a new value of $c$ at the moment of deployment --- or better, model the transition. **(b)** The assumption that the chain is massless fails. An anchor chain can weigh hundreds of kilograms. The tension at the top must support the weight of all the chain below it, so tension varies significantly along the chain's length. **(c)** The assumption that $g$ is constant fails. At 30 km altitude, $g$ has decreased to about $9.7$ m/s$^2$ --- a 1% change. For precise measurements (and weather balloon work often requires precision), this matters. **(d)** The assumption that friction is speed-independent fails. At high speeds on a wet surface, hydroplaning can occur: a layer of water lifts the tire off the road surface, dramatically reducing friction. The friction coefficient becomes a strong function of speed, tire tread design, and water depth.

Layer 3: Structure -- The Value of Wrong Models

Problem 3. A student says: "If every force model is an approximation, and they all fail somewhere, why don't we just use the more accurate model from the start? Why teach $f_k = \mu_k N$ if it's wrong?"

Write a response to this student. Address at least two reasons why simplified models are valuable despite their limitations.

Check your answer

Here are several reasons (any two would be a strong answer): **1. Simplified models are solvable.** The equation $f_k = \mu_k N$ is a constant that drops cleanly into $\sum F = ma$. A speed-dependent friction model $f_k(v) = \mu_k(v) N$ turns the problem into a more complex differential equation that may require numerical methods. You need the simpler model to build understanding before tackling the harder one. **2. Simplified models reveal the dominant physics.** In most everyday situations, the constant-friction model captures 90% or more of the relevant behavior. The corrections are small. Using a more complex model would add computational burden without changing the answer in any meaningful way. Physics is about identifying what matters most, not about including every possible effect. **3. Simplified models provide a baseline for comparison.** You cannot identify a deviation if you do not know what the "normal" prediction is. When a real system behaves differently from the simplified model's prediction, the *pattern of deviation* tells you what physics is missing. The simple model is the reference against which you measure new effects. **4. You cannot understand the complex model without the simple one.** A speed-dependent friction coefficient $\mu_k(v)$ is a modification of the constant model. The concept of "friction coefficient" and its relationship to normal force come from the simple model. You learn to walk before you run.

Layer 4: Creation -- Propose a Better Model

Problem 4. The standard kinetic friction model says $f_k = \mu_k N$, with no speed dependence. Propose a modification that accounts for speed. Your proposal should include:

(a) A modified formula (it does not need to be correct for all materials --- just plausible)

(b) A physical justification for why speed might affect friction

(c) What experimental data you would need to test your model and determine its parameters

Check your answer

Here is one reasonable proposal (many others are possible): **(a)** A modified formula: $$f_k = \mu_k(v)\, N = \left(\mu_0 - \alpha v\right) N$$ where $\mu_0$ is the friction coefficient at low speed and $\alpha$ is a positive constant that captures the decrease in friction with speed. This is a linear model --- the simplest possible speed dependence. Alternatively, you could propose $f_k = \mu_0 N e^{-v/v_0}$, where $v_0$ is a characteristic speed scale at which friction begins to decrease significantly. This model has the advantage of keeping friction positive at all speeds. **(b)** Physical justification: At higher speeds, microscopic surface contacts (asperities) have less time to form strong bonds before they are sheared apart. At low speeds, the surfaces can "settle" into each other, creating more interlocking and adhesion. At high speeds, the contacts are brief and shallow, reducing the effective friction. For lubricated surfaces, higher speed can also create a thicker fluid film between the surfaces (hydrodynamic lubrication), further reducing friction. **(c)** Experimental data needed: - Measure the friction force on a block sliding at several controlled speeds (e.g., 0.1, 0.5, 1, 2, 5, 10, 20 m/s) while keeping the normal force constant - Plot $f_k$ vs. $v$ and check whether the relationship is linear, exponential, or something else - Repeat for different normal forces to verify that the speed dependence and normal-force dependence are separable (i.e., that the model $f_k = \mu_k(v) N$ is the right *structure*) - Determine $\mu_0$ from the low-speed limit and $\alpha$ (or $v_0$) from the slope of the $f_k$-vs-$v$ curve The most important test: does the model predict the friction force at a speed you did *not* use to fit the parameters? If so, the model has predictive power. If not, you need a different functional form.

Reflection

Which force model used in this course are you least confident about --- and why?

Think about the models you have encountered: constant gravity, ideal tension, friction as $\mu N$, drag as $bv$ or $cv^2$, rigid normal forces. Which one feels most fragile? Which one would you most want to test experimentally before trusting it in a real application? There is no wrong answer. The point is to develop the habit of evaluating your own confidence in the tools you use.

Chapter Summary: Forces and Newtonian Kinetics

This chapter answered the question that Chapters 1 through 4 left open: what causes motion to change? The answer is force, and the law that connects force to motion change is Newton's second law, $\sum \vec{F} = m\vec{a}$.

Here is what we built, section by section:

Section 5.1: Free-body diagrams. Every force arises from an interaction with another object. The free-body diagram isolates one object and shows only the forces acting on it. No reliable equation can be written without this diagram.

Section 5.2: Newton's laws. First law: zero net force means constant velocity. Second law: net force produces acceleration proportional to force and inversely proportional to mass. Third law: forces come in equal-and-opposite interaction pairs. These three laws are the foundation of all classical dynamics.

Section 5.3: The equation $\sum \vec{F} = m\vec{a}$. Mass is resistance to acceleration. The working equation of dynamics is Newton's second law in component form. Force does not cause motion --- it causes change in motion.

Section 5.4: Common force models. Weight ($mg$), tension, normal force, friction ($f \leq \mu N$), and drag ($bv$ or $cv^2$) --- each has a physical origin, a mathematical form, and conditions under which it applies.

Section 5.5: Coupled systems. When objects are connected, constraint forces enforce kinematic relationships. Internal forces cancel in whole-system analysis; they appear as unknowns in individual-object analysis.

Section 5.6: Equilibrium. Equilibrium is $\sum \vec{F} = \vec{0}$, which is $\vec{a} = \vec{0}$ --- a special case of the second law, not a separate topic.

Section 5.7: Modeling limits. Every force model is an approximation with a domain of validity. The skill of the physicist is knowing when the model applies and recognizing when it fails. The assumptions behind the model define its limits. The pattern of failure reveals the physics that the model was ignoring.

The hero concept of this chapter is Newton's second law: the bridge between interactions (forces) and motion change (acceleration). Everything in dynamics --- from a block on a ramp to a satellite in orbit --- flows from this single equation and the force models that feed into it. But the equation is only as good as the models that supply the forces, and those models, as this final section showed, are always approximations with boundaries.

Chapter-End Retrieval

Close your notes. Answer these from memory.

1. What are Newton's three laws of motion? State each one in your own words.

2. What is a free-body diagram? Why is it essential for solving dynamics problems?

3. What is the difference between mass and weight? One is a property of the object; the other is a force. Which is which?

4. Is the normal force always equal to $mg$? Give an example where it is not.

5. What is the difference between static and kinetic friction? Why is friction modeled as an inequality ($f_s \leq \mu_s N$) rather than an equation?

6. What does it mean for a force model to have a "domain of validity"? Give an example of a model being used outside its domain.

After you have attempted all six, check your answers against the chapter summary above.

Looking Ahead

Chapter 5 gave you the conceptual framework: forces cause acceleration, Newton's second law connects them, and free-body diagrams organize the reasoning. But if you are honest, most of your time so far has been spent understanding the ideas rather than solving complex problems. The machinery is assembled. Now it needs to be used.

Chapter 6 takes the tools of this chapter into messy, multi-step problems --- blocks on inclines connected by ropes, friction that determines whether a system moves or stays still, circular motion that demands inward force. The physics will not change. Newton's laws are the same. The force models are the same. What changes is the complexity of the setup and the strategic decisions you need to make: which object to isolate, which coordinates to choose, which forces matter and which can be ignored.

The lesson of Section 5.7 will become immediately practical in Chapter 6. Every problem you solve will require you to make modeling choices --- to decide which forces to include, which to neglect, and whether the standard models apply. The meta-skill you developed here, evaluating models rather than blindly applying them, will be the difference between setting up problems correctly and getting lost in equations that do not match the physics.