ARCHETYPE: Stacked Blocks & Friction Between

Pull a tablecloth slowly and the dishes come with it. Snap it fast and the dishes stay put while the cloth slides out from under them. Same cloth, same dishes, same friction — two completely different outcomes. Every stacked-block exam problem lives inside that trick, and by the end of this article you will be able to compute exactly how fast is "fast." The setup is minimal: a block rides on a slab, friction between them, and a force applied to one or the other. What makes it the most conceptually loaded archetype in the Newton gym is that here friction is the engine, not the brake. Whichever body you don't push has exactly one horizontal force available to make it move: the friction from its partner. The passive body is dragged along by friction — and friction can only drag so hard. Below the threshold, the stack moves as one; above it, the interface gives way and the two bodies live separate dynamical lives with kinetic friction trading between them. NCERT plants the seed with its block-on-a-trolley exercise; NEET tests the threshold; JEE-Main loves the wicked asymmetry (pushing the slab allows several times more force than pushing the block — with the same interface!); JEE-Advanced runs the slipping regime until the block falls off the back of the slab. All of it is one decision tree: assume together, compute the friction demanded, compare with the budget, and split the system only if the interface defaults. This article drills that tree, the third-law bookkeeping that powers it, and the traps that make "who drags whom" the sharpest friction question on any paper.

Key takeaways

Stacked blocks (mass $m$ on slab $M$ ) move together as long as the friction between them can supply what the passive body needs: demand $f = (\text{passive mass}) \times a$ , budget $\mu_s N$ with $N = mg$ — only the top block's weight presses the interface.
The decision tree: assume together → $a = F/(M+m)$ → compute demanded friction → compare with $\mu_s mg$ . Demand ≤ budget: together. Demand > budget: slipping, switch to kinetic friction and separate FBDs.
Pushing the slab: friction drags the light top block — threshold $F_{\max} = \mu_s(M+m)g$ . Pushing the block: friction must drag the heavy slab — threshold collapses to $\mu_s mg(M+m)/M$ . Same interface, very different limits.
In the slipping regime the friction pair is kinetic: the pushed body runs at its own $a$ , the passive body is driven by $\mu_k mg$ alone, and the relative acceleration times the slab length sets when the block falls off.
Friction here drives: on the passive body it is the only forward force. "Friction opposes relative sliding," never simply "motion."
Key idea: who drags whom — the passive body rides on friction alone, so the stack's fate is the comparison (passive mass) × a versus μ_s mg.

The core insight

A stacked pair shares one acceleration only while the interface friction can supply the passive body's demand $(m_{\text{passive}} \times a) \le \mu_s m_{\text{top}} g$ ; beyond that threshold the surfaces slip, kinetic friction $\mu_k m_{\text{top}} g$ becomes a known force on each body, and the two are solved separately.

Where this lives today

Logistics is stacked-block engineering at fleet scale. A loaded pallet on a forklift, cartons on a conveyor tray, your groceries on the car seat — each has a friction budget $\mu_s mg$ and a maximum safe acceleration $\mu_s g$ , which is why airport baggage tractors and warehouse AGVs carry explicit acceleration caps in their firmware (typically $\sim 2\ \mathrm{m/s^2}$ , a $\mu_s$ of 0.2 with margin). Railways live the inverse problem: a locomotive's driving wheel is the pushed body and the train is the passenger, with steel-on-steel adhesion ( $\mu \approx 0.3$ , dropping with rain) capping tractive force — sanding systems exist purely to raise the interface budget. And earthquake engineering deliberately inverts the archetype: base-isolated buildings sit on low-friction bearings so that when the ground (the slab) lurches, the structure (the block) is dragged only weakly — the plateau branch of the slipping graph, exploited to keep hospitals standing. The tablecloth trick, at building scale, on purpose.

The setup

Geometry first, because in an archetype the geometry is the physics. A block of mass $m$ rests on a slab of mass $M$ ; the slab rests on a smooth floor (the base case — a rough floor is a bolt-on). The coefficient between block and slab is $\mu_s$ (static) / $\mu_k$ (kinetic). A horizontal force $F$ is applied — to the slab, or to the block, and the two versions are different problems.

Figure

Block on slab — the friction pair that carries the passive body, drawn as the third-law pair it is

Wait — the friction arrow on the block points forward, the direction of motion? Yes, and that arrow is the whole archetype. Audit the block: gravity down, normal up, and one horizontal candidate — friction from the slab. If the block accelerates at all, friction did it. Friction is a brake only for bodies being pushed; for passengers, it is the engine. (Slogan worth engraving: friction opposes relative sliding between the surfaces, not "motion." When the slab tries to slide forward under the block, friction on the block acts forward to prevent that relative slide.) Two pieces of bookkeeping power everything downstream:

The interface is pressed only by what sits on it. Perpendicular equilibrium of the block: $N_{\text{interface}} = mg$ . The slab's own weight routes around the interface, through the slab's bottom to the floor ( $N_{\text{floor}} = (M+m)g$ ). So the friction budget between the bodies is $\mu_s mg$ — the top mass only — no matter how heavy the slab. This single line decides most MCQs.

The friction is a third-law pair. Whatever friction drags the passenger forward, its equal-and-opposite twin drags the pushed body backward. Forces internal to the stack — which is why the together-regime system line $F = (M+m)a$ is legal, and why the pair must both appear the moment you separate the FBDs.

The master solution pattern — a decision tree

Step 1 — assume together. One body: $a = \dfrac{F}{M+m}$ .

Step 2 — price the demand. The passive body (the one not pushed) needs $f_{\text{demand}} = m_{\text{passive}}\, a$ — friction is its only engine.

Step 3 — compare with the budget.

f_{\text{demand}} \le \mu_s\, m g \quad\Rightarrow\quad \text{together (assumption holds)} \qquad (1)

where: $\mu_s mg$ — the interface ceiling, priced by the top block's weight always (it is the interface's normal force, whichever body is pushed). If the inequality holds, you are done: the actual static friction equals the demand, not the ceiling.

Step 4 — if the budget fails, split. The surfaces slip; kinetic friction $f_k = \mu_k mg$ becomes a known force — forward on whichever body lags, backward on whichever leads:

a_{\text{pushed}} = \frac{F - \mu_k mg}{m_{\text{pushed}}}, \qquad a_{\text{passive}} = \frac{\mu_k mg}{m_{\text{passive}}} \qquad (2)

where: the passive body's acceleration is now fixed — $\mu_k mg$ divided by its mass — no matter how monstrously $F$ grows; extra force accelerates only the pushed body and widens the slip.

Derivation AThe together threshold — and why pushing the slab is so much safer

We derive the maximum force for the stack to move as one, first with $F$ on the slab, then with $F$ on the block, exposing the asymmetry examiners adore. Strategy: the decision tree with the inequality run to its edge. Watch a single subtlety: the interface budget $\mu_s mg$ never changes — only the demand changes sides.

Step 1. Push the slab. Together: $a = \dfrac{F}{M+m}$ . The passive body is the light block $m$ : demand $= ma = \dfrac{mF}{M+m}$ .

Step 2. Threshold — demand meets budget:

\frac{mF_{\max}}{M+m} = \mu_s m g \;\;\Rightarrow\;\; F_{\max} = \mu_s (M+m)g

Why: the $m$ cancels — the threshold looks like "friction coefficient times the WHOLE stack's weight," a formula so clean it tempts you to use it in the next case. Resist.

Step 3. Push the block. Together: same $a = \dfrac{F}{M+m}$ . But the passive body is now the heavy slab $M$ : demand $= Ma = \dfrac{MF}{M+m}$ , while the budget is still $\mu_s mg$ (interface unchanged!).

Step 4. Threshold:

\frac{MF'_{\max}}{M+m} = \mu_s m g \;\;\Rightarrow\;\; F'_{\max} = \mu_s m g\,\frac{M+m}{M}

For $M \gg m$ this tends to just $\mu_s mg$ — tiny. Same interface, same coefficient: pushing the slab tolerates $\dfrac{F_{\max}}{F'_{\max}} = \dfrac{M}{m}$ times more force. Verified: limits — equal masses give thresholds $2\mu_s mg$ vs $2\mu_s mg$ … equal, as symmetry demands; $M \to \infty$ pushed-slab threshold diverges (an infinite slab accelerates at zero — no demand), pushed-block threshold saturates at $\mu_s mg$ . Numbers: $m = 1$ kg, $M = 4$ kg, $\mu_s = 0.3$ , $g = 10$ : push slab 15 N, push block 3.75 N — the factor $M/m = 4$ ✓.

Go deeper · Derivation B · 1 minThe slipping regime — two lives, one interface, and the fall off the back

Derivation BThe slipping regime — two lives, one interface, and the fall off the back

We solve the stack after the interface defaults: force $F$ (large) applied to the top block, which now slides forward on the slab. We find both accelerations and the time for the block to slide off a slab of length $L$ — the JEE-Advanced finisher. Strategy: separate FBDs with the kinetic pair as known forces, then work in the relative coordinate. Watch the pair's directions: the block slides forward relative to the slab, so kinetic friction on the block acts backward, and its twin on the slab acts forward.

Step 1. Block ( $F$ forward, kinetic friction $\mu_k mg$ backward):

a_1 = \frac{F - \mu_k m g}{m}

Step 2. Slab (driven only by the friction twin, forward):

a_2 = \frac{\mu_k m g}{M}

Why: the slab has stopped caring about $F$ entirely — its only messenger from the block is the kinetic friction, now a fixed number. Pushing the block harder speeds the BLOCK up, not the slab: extra force goes into widening the slip.

Step 3. Relative acceleration of block over slab:

a_{\text{rel}} = a_1 - a_2 = \frac{F - \mu_k mg}{m} - \frac{\mu_k mg}{M}

Step 4. Starting from rest with the block a distance $L$ from the slab's rear edge, the slip grows as $\tfrac{1}{2}a_{\text{rel}}t^2$ ; the block leaves the slab when the slip equals $L$ :

t = \sqrt{\frac{2L}{a_{\text{rel}}}}

Verified: consistency — at the threshold force the two accelerations merge ( $a_1 = a_2$ ) with $\mu_k \to \mu_s$ , closing the tree. Numbers: $m = 1$ , $M = 4$ kg, $\mu_k = 0.2$ , $F = 10$ N, $L = 1.5$ m, $g = 10$ : $a_1 = 8$ , $a_2 = 0.5\ \mathrm{m/s^2}$ , $a_{\text{rel}} = 7.5$ , $t = \sqrt{0.4} = 0.63$ s.

The regimes at a glance

Figure

Acceleration versus applied force — one line while together, two branches after the interface defaults

question	move
"do they move together?"	tree Steps 1–3: demand vs $\mu_s mg$
max $F$ on the slab	$\mu_s(M+m)g$
max $F$ on the block	$\mu_s mg(M+m)/M$
actual friction while together	the demand $m_{\text{passive}}a$ — NOT the ceiling
accelerations after slipping	eq. (2) with $\mu_k$
when does the block fall off	relative coordinate: $t = \sqrt{2L/a_{\text{rel}}}$

Limiting cases

case	result	meaning
$\mu = 0$	slab shoots out, block stays	the tablecloth trick, exactly
$\mu \to \infty$	welded stack, never slips	one rigid body at $F/(M+m)$
$M \gg m$ , push slab	huge threshold	massive slab barely accelerates; tiny demand on the interface
$M \gg m$ , push block	threshold $\to \mu_s mg$	dragging a mountain by a napkin's friction
$F \to \infty$ in slip regime	$a_{\text{passive}}$ plateaus at $\mu_k mg/m_{\text{passive}}$	the passenger can't hear the extra force
rough floor added	subtract floor friction $\mu_f(M+m)g$ from the system line	bolt-on, same tree

Common trap"Friction always opposes motion, so it must point backward on the block"

On the passive block, friction points forward — in the direction of motion — and it is the only reason the block moves at all. The slogan "friction opposes motion" is a corrupted version of the real law: friction opposes relative sliding between the two touching surfaces (actual sliding for kinetic, would-be sliding for static). When the slab is pushed, the slab tends to slide forward underneath the block — equivalently, the block tends to slide backward relative to the slab — and friction on the block acts forward to fight that relative slide. Draw the third-law twin and the books balance: forward $f$ on the block, backward $f$ on the slab, the pair transmitting the push upward through the interface. Every stacked-blocks FBD error traces to the corrupted slogan; every correct one starts with the question "which way would this surface slide over its partner if friction quit?" Ask it surface by surface, and conveyor belts, clutch plates, and this whole archetype fall into line.

Worked examples

1 · Board level — the NCERT trolley test (CBSE pattern, 3 marks)

A $15$ kg block sits on a long trolley; $\mu_s = 0.18$ between them. The trolley accelerates from rest at $0.50\ \mathrm{m/s^2}$ . Take $g = 10\ \mathrm{m/s^2}$ . Does the block slip? What friction force acts on it?

Step 1. Assume together: the block must be dragged at $a = 0.50\ \mathrm{m/s^2}$ ; demand $f = ma = 15 \times 0.50 = 7.5$ N.

Step 2. Budget: $\mu_s mg = 0.18 \times 15 \times 10 = 27$ N.

Step 3. Demand $7.5 \le 27$ : no slip — the stack rides together.

Step 4. The actual friction is the demand, $\boxed{7.5\ \mathrm{N}}$ , pointing forward (it is what accelerates the block) — not 27 N. Quoting the ceiling when the question asks the force is the classic lost mark; the ceiling is a limit, not a reading.

2 · Board level — maximum acceleration of the base (NEET/AP pattern)

A $2.0$ kg box rests on a flatbed cart of $6.0$ kg; $\mu_s = 0.25$ , floor smooth, $g = 10\ \mathrm{m/s^2}$ . Find the maximum force on the cart for the box to ride without sliding, and the friction on the box at that maximum.

Step 1. The box's engine is friction, ceiling $\mu_s mg = 0.25 \times 2.0 \times 10 = 5.0$ N ⇒ its maximum possible acceleration is $a_{\max} = \mu_s g = 2.5\ \mathrm{m/s^2}$ .

Step 2. The stack must not out-accelerate the box: $F_{\max} = (M+m)a_{\max} = 8.0 \times 2.5 = \boxed{20\ \mathrm{N}}$ .

Step 3. At the verge the friction on the box sits at its ceiling: $\boxed{5.0\ \mathrm{N}}$ , forward.

Step 4. Sanity via the formula: $F_{\max} = \mu_s(M+m)g = 0.25 \times 8.0 \times 10 = 20$ N ✓. Note the two-step logic — ceiling on the passenger first, then scale up to the stack — it survives any dressing (trolleys, flatbeds, books on car seats).

3 · Competitive level — the asymmetry, both ways (JEE-Main pattern, ~3 minutes)

A $1.0$ kg block lies on a $4.0$ kg slab on a smooth floor; $\mu_s = 0.30$ , $g = 10\ \mathrm{m/s^2}$ . Find the maximum horizontal force for no slipping when applied (a) to the slab, (b) to the block. (c) Explain the difference in one sentence.

Step 1. Interface budget (both cases): $\mu_s mg = 0.30 \times 1.0 \times 10 = 3.0$ N.

Step 2. (a) Push the slab — friction drags the 1 kg block: its max acceleration is $\mu_s g = 3.0\ \mathrm{m/s^2}$ , so $F_{\max} = (M+m)(3.0) = 5.0 \times 3.0 = \boxed{15\ \mathrm{N}}$ .

Step 3. (b) Push the block — friction must drag the 4 kg slab: slab's max acceleration $= 3.0/4.0 = 0.75\ \mathrm{m/s^2}$ , so $F'_{\max} = 5.0 \times 0.75 = \boxed{3.75\ \mathrm{N}}$ .

Step 4. (c) The budget is fixed by the top block's weight, but the demand is the passive body's mass times $a$ — friction dragging the heavy slab maxes out four times sooner than friction dragging the light block. Timing note: both thresholds come from one ritual — passenger's max acceleration = (budget)/(passenger mass), then × total mass. Under 3 minutes with the ritual; the trap is reusing $\mu_s(M+m)g$ for case (b), which quadruple-counts the interface's strength. The MCQ bait 15 N will be among case (b)'s options, guaranteed.

4 · Competitive level — JEE-Advanced stretch: sliding off the back (~4.5 minutes)

A $1.0$ kg block sits at the front end of a $4.0$ kg slab of length $L = 1.5$ m on a smooth floor; $\mu_k = 0.20$ between them ( $\mu_s$ barely higher). A constant $10$ N force is applied to the block. Take $g = 10\ \mathrm{m/s^2}$ . Show the pair slips, find both accelerations, and the time for the block to slide off the slab.

Step 1. Slip check: together would need $a = 10/5.0 = 2.0\ \mathrm{m/s^2}$ ; demand on the interface = $Ma = 8.0$ N; budget $\approx \mu mg = 2.0$ N. Demand ≫ budget ⇒ slips from the start.

Step 2. Block: $a_1 = \dfrac{F - \mu_k mg}{m} = \dfrac{10 - 2.0}{1.0} = 8.0\ \mathrm{m/s^2}$ . Slab (friction twin drives it): $a_2 = \dfrac{\mu_k mg}{M} = \dfrac{2.0}{4.0} = 0.50\ \mathrm{m/s^2}$ .

Step 3. Relative acceleration: $a_{\text{rel}} = 8.0 - 0.50 = 7.5\ \mathrm{m/s^2}$ . The block must gain $L = 1.5$ m on the slab:

t = \sqrt{\frac{2L}{a_{\text{rel}}}} = \sqrt{\frac{3.0}{7.5}} = \sqrt{0.4} = \boxed{0.63\ \mathrm{s}}

Step 4. Optional finisher examiners bolt on: speeds at separation — block $v_1 = 8.0 \times 0.63 = 5.1$ m/s, slab $v_2 = 0.32$ m/s; from separation onward the slab coasts (smooth floor) and the block, still pushed, accelerates at $10\ \mathrm{m/s^2}$ . Timing note: the relative coordinate is the entire trick — one equation $\tfrac{1}{2}a_{\text{rel}}t^2 = L$ replaces tracking two positions. Do the slip check FIRST (ten seconds) or you'll solve a together-problem that never existed. Budget 4.5 min. Bonus concept: the slab's $0.5\ \mathrm{m/s^2}$ is friction-driven — the only force moving it is the block's kinetic friction twin.

Exam traps

Ceiling quoted as the force. While the stack rides together, friction equals the demand ( $m_{\text{passive}}a$ ), not $\mu_s mg$ . Example 1's 7.5-vs-27 N is the canonical test; the ceiling appears only at the verge or in the slip regime.

Interface pressed by the whole stack. $N_{\text{interface}} = mg$ — the top block's weight only. The slab's weight goes to the floor. Using $(M+m)g$ for the interface budget silently multiplies thresholds by the wrong factor.

One threshold formula for both push cases. $\mu_s(M+m)g$ belongs to pushing the slab. Pushing the block collapses the limit by the factor $m/M$ — Derivation A. Where the force lands is physics, not decoration.

Friction drawn backward on the passenger. The passive body's friction points forward — it is the engine (see the pitfall box). Backward friction on the passenger leaves it with no mover and an impossible FBD.

Missing the third-law twin. The moment you split FBDs, the kinetic friction must appear on both bodies, equal and opposite. A slab FBD without the forward twin predicts a slab that never moves — visibly wrong in the tablecloth limit.

Slip check skipped. Applying together-formulas to a slipping stack (or vice versa) is the archetype's structural error. Ten seconds: demand vs budget, then commit to a branch.

Which $g$ . CBSE/NEET as declared (usually 10 in JEE/NEET; NCERT exercises use 10 here); AP allows 10. Thresholds scale linearly with $g$ .

Origins · 1 minCharles-Augustin de Coulomb and how this was discovered

Charles-Augustin de Coulomb

1736–1806 · Paris Academy of Sciences

Coulomb met friction as a military engineer dragging cannon and capstans through the dockyards of Rochefort, and answered the Academy's 1781 prize question with the memoir that still governs this archetype: friction is proportional to the normal load, largely independent of contact area, and stronger at rest than in motion — the $\mu_s > \mu_k$ gap on which every tablecloth trick and every slip threshold in this article turns. His laws were unapologetically empirical, distilled from wooden sledges, tallow, and hemp rope rather than theory; two centuries of tribology have added atoms and asperities underneath without displacing the equations. The same man, with the same torsion balance he built to measure tiny forces, then quantified the electrostatic inverse-square law — putting his name simultaneously on the force that holds your phone on the dashboard and the unit of the charge inside it.

History

1699

— Amontons rediscovers the load-proportionality of friction; area independence scandalizes the Academy.

1781

— Coulomb's prize memoir separates static from kinetic friction and fixes the laws used here.

1900

s — locomotive engineering formalizes "adhesion": driving wheels are the stacked-block problem with steel on steel.

1950

— Bowden and Tabor explain

\mu

via micro-contact asperities — why the budget tracks load, not area.

1990

s — seismic base isolation puts entire buildings on low-μ bearings: the slab slides so the block (the building) barely feels the quake.

Going further

Texts. NCERT Physics Class 11, Ch. 4 exercises (block-on-trolley, box-in-truck) · HC Verma, Concepts of Physics Vol. 1, Ch. 6 · DC Pandey Mechanics 1 (the slipping-regime problem set).

Provenance. All results are standard. A verifier should recompute: Derivation A's thresholds $\mu_s(M+m)g$ vs $\mu_s mg(M+m)/M$ and the 15 / 3.75 N pair; Example 1's 7.5 N vs 27 N; Example 2's 20 N; Example 4's $a_1 = 8.0$ , $a_2 = 0.5\ \mathrm{m/s^2}$ , $t = \sqrt{0.4} = 0.63$ s.

Prerequisites. Static Friction · Kinetic & Rolling Friction · Newton's Third Law · Connected Blocks, Strings & Constraints. Leads to. Conveyor Belts · Topple or Slide? (rotation). Deep dive (UG). Friction Forces (UG treatment, two-block slip threshold).

Why is the interface normal force just mg and not the whole stack's weight?

Follow the load path. The block's weight presses the interface; the interface answers with $N = mg$ (the block's perpendicular equilibrium demands exactly that). The slab's own weight never crosses the interface — it is born in the slab and exits through the slab's bottom into the floor, which is why the floor feels $(M+m)g$ . Each surface carries only what sits above it — the same cut-and-count logic as tensions in a hanging rope, rotated horizontal. Since the friction budget is $\mu_s N$ , the interface's grip is priced by the passenger's weight alone, however massive the vehicle.

How can friction drive something forward? I thought it only ever resists.

It only ever resists relative sliding between the touching surfaces — and preventing relative sliding frequently requires pushing a body forward. Your own walking works exactly this way: your foot pushes the ground backward, and the ground's friction pushes you forward; a car's driven wheel does the same. The stacked block is the cleanest textbook case: the slab slides (or tends to slide) forward beneath the block, so friction on the block acts forward to suppress the relative slip — and that forward force is the block's entire engine. Resistance at the interface, propulsion for the body.

Why does it matter which body I push? The interface has one μ and one normal force either way.

Because the interface's job changes. The budget $\mu_s mg$ is indeed fixed — but friction's task is to drag along whichever body you did NOT push, and dragging a heavy slab demands far more force than dragging a light block at the same acceleration. Push the slab: friction hauls little $m$ , demand $= mF/(M+m)$ , threshold high. Push the block: friction hauls big $M$ , demand $= MF/(M+m)$ , threshold low by the factor $m/M$ . Fixed budget, movable burden — the asymmetry is entirely in who rides on friction.

What is the actual friction force when the stack moves together well below the threshold?

Exactly the demand: $f = m_{\text{passive}} \times a$ , computed from the shared acceleration — static friction is self-adjusting and supplies no more than the job requires. The ceiling $\mu_s mg$ is a capability, touched only at the verge. This is the static-friction module's demand-versus-budget law wearing stacked clothes, and it is why Example 1's answer is 7.5 N while the ceiling sits at 27 N. If a question's given acceleration implies a demand above the ceiling, the together-assumption was false and you must re-solve in the slip regime.

In the slipping case, why doesn't the slab speed up when I push the block harder?

Because the slab's only connection to your push is the kinetic friction across the interface, and kinetic friction is a fixed number: $\mu_k mg$ , blind to $F$ . Once slipping, the interface transmits that and nothing more — the slab plateaus at $a_2 = \mu_k mg/M$ while every extra newton of $F$ goes into accelerating the block and widening the slip. The graph of the regimes shows it as the flat branch. (Pedantic footnote with real content: $\mu_k mg$ is fixed while the block stays on the slab and $N = mg$ — tilt the setup or add vertical forces and the plateau re-prices.)

Does the tablecloth trick really work by this physics, and why does speed matter?

Entirely. Cloth = slab (pushed), dishes = passengers riding on a small $\mu_k$ (smooth cloth, glazed china). While you pull, kinetic friction does drag the dishes — at $a = \mu_k g \approx 1\text{–}2\ \mathrm{m/s^2}$ , regardless of how hard you yank (the plateau again!). Speed matters because it shrinks the time the interface exists: pull the cloth out in 0.1 s and the dishes gain only $v = \mu_k g\,t \approx 0.15$ m/s and drift a centimetre; pull slowly over 2 s and the same acceleration acts forty times longer, marching the dishes off the table with the cloth. The trick is not beating friction — it is starving friction of time. Physics grade: Example 4 run in reverse, with dinner service as the passive body.