Visual walkthrough — Vectors in ℝⁿ — operations, geometric interpretation

4.5.1 · D2 · Maths › Linear Algebra (Full) › Vectors in ℝⁿ — operations, geometric interpretation

Recall Quick refresher:

$\cos \theta$ kya hota hai? Ek right triangle lo jisme ek angle $\theta$ ho. Tab $\cos \theta =\frac{\text{adjacent leg}}{\text{hypotenuse}},$ $\theta$ ke paas wali side ka ratio slanted side se. Yeh smoothly slide karta hai $\cos 0^{\circ }=1$ se (arrows lined up), $\cos 90^{\circ }=0$ se hote hue (right angle), aur $\cos 180^{\circ }=-1$ tak (opposite direction mein point karte hue). Woh single number yehi hai jisse "kitne aligned hain?" ko arithmetic mein convert kiya jaata hai — isliye yeh Law of Cosines mein neeche dikhta hai.

---

Step 1 — Arrow actually kya hota hai?

KYA. Hum do arrows draw karte hain, $\mathbf{u}$ aur $\mathbf{v}$, dono origin se start hote hue, flat plane ${\mathbb{R}}^2$ mein.

KYUN. Angle ke baare mein sab kuch in do arrows ke beech rehta hai, isliye pehle hume exactly pin down karna hoga ki arrow kya hai: ek horizontal amount aur ek vertical amount, do axes se padhe hue.

PICTURE. Figure dekho. Burnt-orange arrow $\mathbf{u}$ hai; teal arrow $\mathbf{v}$ hai. Woh angle jo hum samajhna chahte hain — plum wedge jis par $\theta$ ("theta", Greek letter jo hum unknown angle ke liye use karte hain) likha hai — do arrows ke beech ki opening hai.

Yahan har symbol ka matlab hai:

• $u_1$ ::: kitna door $\mathbf{u}$ east mein jaata hai (uska first coordinate).
• $u_2$ ::: kitna door $\mathbf{u}$ north mein jaata hai (uska second coordinate).
• $\theta$ ::: do arrows ke beech ka angle — woh cheez jo hum dhundh rahe hain.

---

Step 2 — Length Pythagoras hai, ek axis at a time

KYA. Hum measure karte hain ki arrow kitna lamba hai, use ek right triangle ke slanted side (hypotenuse) ki tarah treat karke jiske do legs uske coordinates hain.

KYUN yeh tool — Pythagoras? Kyunki east-leg aur north-leg ek perfect right angle par milte hain (axes perpendicular hain), aur Pythagoras exactly woh theorem hai jo do perpendicular legs ko slant ki length mein convert karta hai. Koi aur tool sideways+upward ko straight-line distance mein convert nahi karta.

PICTURE. Dashed legs $v_1$ (horizontal) aur $v_2$ (vertical) ek L banate hain; arrow ise ek right triangle mein close karta hai. Legs ko square karke add karna, phir square root lena, arrow ki asli length deta hai.

$\sqrt{v_1^2+v_2^2}$ ke andar term by term:

• $v_1^2$ ::: horizontal leg, squared — diagonal mein uska contribution.
• $v_2^2$ ::: vertical leg, squared — diagonal mein uska contribution.
• woh ​ ::: squaring ko undo karta hai taaki hum ek honest length par wapas aayen, area par nahi.

Recall End mein square root kyun?

$v_1^2+v_2^2$ ke units area hain (length times length). Length one-dimensional hai, isliye hum square root lete hain taaki ek plain distance par wapas aa sakein.

---

Step 3 — Dot product: hamara raw material

KYA. Hum ek purely arithmetic recipe define karte hain: east coordinates ko pair up karo, north coordinates ko pair up karo, har pair ko multiply karo, add karo. Abhi yeh ek random bookkeeping jaisa lagta hai.

KYUN ABHI introduce karte hain? Kyunki yeh woh object hai jiska geometric meaning hum discover karna chahte hain. Hum ise pehle table par rakhte hain taaki baad mein, jab Pythagoras hume numbers ka yahi exact combination de, hum ise ek baar mein pehchaan sakein.

PICTURE. Figure do "matching-coordinate" multiplications ko coloured brackets ki tarah dikhata hai: $u_1{v}_1$ (east pair) aur $u_2{v}_2$ (north pair), box mein ek number mein summed.

Definition se seedha ek chhota lekin crucial special case: $\mathbf{v}\cdot \mathbf{v}=v_1{v}_1+v_2{v}_2=v_1^2+v_2^2=\Vert \mathbf{v}{\Vert }^2.$ Toh ek vector khud se dotted apni length squared hoti hai. Ise yaad rakho — hum ise Step 6 mein use karte hain.

• $u_1{v}_1$ ::: $\mathbf{u}$ ka east-coordinate times $\mathbf{v}$ ka east-coordinate.
• $u_2{v}_2$ ::: $\mathbf{u}$ ka north-coordinate times $\mathbf{v}$ ka north-coordinate.
• woh $+$ ::: do axis-contributions ko ek number mein fuse karta hai.

---

Step 4 — Do arrows se ek triangle banao

KYA. Hum woh arrow draw karte hain jo $\mathbf{v}$ ke tip se $\mathbf{u}$ ke tip tak jaata hai. Yeh closing side exactly $\mathbf{u}-\mathbf{v}$ hai (coordinates subtract karo: $(u_1-v_1,u_2-v_2)$).

SUBTRACT KYUN, aur yeh shape kyun? Angle $\theta$ is triangle ke andar baitha hai, do arrows ke beech. Koi bhi law jo $\theta$ mention karta hai usmein triangle ke teeno sides involve hone chahiye. Do arrows do sides hain; difference $\mathbf{u}-\mathbf{v}$ woh missing teesra side hai jo "angle ko opposite se dekhta hai."

PICTURE. Plum side $\mathbf{u}-\mathbf{v}$ hai; woh origin par angle $\theta$ face karta hai. Do sides origin se aati hain, teesra top ko close karta hai.

• $\mathbf{u}-\mathbf{v}$ ::: $\mathbf{v}$ ke tip se $\mathbf{u}$ ke tip tak arrow; iski length woh side hai jo angle $\theta$ ke opposite hai.
• $\theta$ ::: origin par angle, $\mathbf{u}$ aur $\mathbf{v}$ ke beech wedged, us teesre side ke opposite.

---

Step 5 — Triangle ko Law of Cosines se padho

KYA. Hum is law ko apne triangle par apply karte hain, $a=\Vert \mathbf{u}\Vert$, $b=\Vert \mathbf{v}\Vert$, aur $c=\Vert \mathbf{u}-\mathbf{v}\Vert$ (plum side) ke saath: $\Vert \mathbf{u}-\mathbf{v}{\Vert }^2=\Vert \mathbf{u}{\Vert }^2+\Vert \mathbf{v}{\Vert }^2-2\Vert \mathbf{u}\Vert \Vert \mathbf{v}\Vert \cos \theta .$

LAW OF COSINES KYUN aur plain Pythagoras kyun nahi? Pythagoras sirf tab kaam karta hai jab $\theta =90^{\circ }$ ho. Hamare do arrows kisi bhi direction mein point karte hain. Law of Cosines exactly Pythagoras hai arbitrary angles ke liye repair kiya gaya: extra term $-2ab\cos \theta$ correction hai. Jab $\theta =90^{\circ }$, $\cos 90^{\circ }=0$, woh term vanish ho jaata hai, aur hum ordinary Pythagoras recover karte hain — ek reassuring sanity check. (Yaad karo $\cos \theta$ = adjacent/hypotenuse, upar ke refresher se.)

PICTURE. Figure usi triangle ko dikhata hai jisme har side ko uski length se label kiya gaya hai, aur correction term $-2\Vert \mathbf{u}\Vert \Vert \mathbf{v}\Vert \cos \theta$ ko "angle ka fingerprint" ki tarah highlight kiya gaya hai.

• $\Vert \mathbf{u}-\mathbf{v}{\Vert }^2$ ::: teesre side ki squared length (woh side jo hum agle step mein coordinates se bhi compute kar sakte hain).
• $\Vert \mathbf{u}{\Vert }^2,\Vert \mathbf{v}{\Vert }^2$ ::: do arrows ki squared lengths.
• $-2\Vert \mathbf{u}\Vert \Vert \mathbf{v}\Vert \cos \theta$ ::: ekmatr jagah jahan $\theta$ aata hai — angle ka hook.

---

Step 6 — Usi side ko algebraically compute karo

KYA. Step 3 se yaad karo ki koi bhi vector khud se dotted apni squared length deta hai. Toh $\Vert \mathbf{u}-\mathbf{v}{\Vert }^2=(\mathbf{u}-\mathbf{v})\cdot (\mathbf{u}-\mathbf{v}).$ Is product ko waisi hi expand karo jaise $(x-y)(x-y)=x^2-2xy+y^2$ expand karte hain, lekin dot products ke saath:

= \underbrace{\mathbf u\cdot\mathbf u}_{\|\mathbf u\|^2} \;-\; \underbrace{\mathbf u\cdot\mathbf v + \mathbf v\cdot\mathbf u}_{2\,\mathbf u\cdot\mathbf v} \;+\; \underbrace{\mathbf v\cdot\mathbf v}_{\|\mathbf v\|^2}.$$ Toh $$\|\mathbf u-\mathbf v\|^2 = \|\mathbf u\|^2 - 2\,\mathbf u\cdot\mathbf v + \|\mathbf v\|^2.$$ **FOIL JAISA EXPAND KARNE KI PERMISSION KYUN HAI?** Kyunki dot product **distributive** hai — yeh addition par split ho jaata hai bilkul ordinary multiplication ki tarah — aur **symmetric** hai, toh $\mathbf u\cdot\mathbf v=\mathbf v\cdot\mathbf u$, isliye do middle terms merge ho ke $2\,\mathbf u\cdot\mathbf v$ bante hain. (Dono facts directly coordinate definition $\sum u_iv_i$ se follow hote hain.) **PICTURE.** Figure expansion ko dot-product tiles ki $2\times2$ grid ki tarah dikhata hai (multiplication table ki tarah), jisme do diagonal tiles lengths ban jaati hain aur do off-diagonal tiles $-2\,\mathbf u\cdot\mathbf v$ mein merge ho jaati hain. ![[deepdives/dd-maths-4.5.01-d2-s06.png]] - $\mathbf u\cdot\mathbf u=\|\mathbf u\|^2$ ::: $\mathbf u$ ki squared length (Step 3 ka special case). - $\mathbf v\cdot\mathbf v=\|\mathbf v\|^2$ ::: $\mathbf v$ ki squared length. - $-2\,\mathbf u\cdot\mathbf v$ ::: purely *algebraic* cross-term — dot product yahan surface hota hai. --- ## Step 7 — Dono answers ko equal set karo aur cancel karo **KYA.** Ab hamare paas $\|\mathbf u-\mathbf v\|^2$ do tarike se likha hua hai. Unhe line up karo: $$\underbrace{\|\mathbf u\|^2 - 2\,\mathbf u\cdot\mathbf v + \|\mathbf v\|^2}_{\text{algebra (Step 6)}} \;=\; \underbrace{\|\mathbf u\|^2 + \|\mathbf v\|^2 - 2\|\mathbf u\|\|\mathbf v\|\cos\theta}_{\text{geometry (Step 5)}}$$ **EQUAL KYUN?** Yeh *ek hi* triangle ke *ek hi* side ke do sachche measurements hain. Agar maths consistent hai, toh yeh disagree nahi kar sakte. Ab **dono** sides se $\|\mathbf u\|^2$ aur $\|\mathbf v\|^2$ subtract karo — yeh dono sides par identically appear karte hain, toh yeh vanish ho jaate hain: $$-2\,\mathbf u\cdot\mathbf v \;=\; -2\|\mathbf u\|\|\mathbf v\|\cos\theta.$$ Dono sides ko $-2$ se divide karo: > [!formula] Result, zero se kamaya hua > $$\boxed{\;\mathbf u\cdot\mathbf v \;=\; \|\mathbf u\|\,\|\mathbf v\|\,\cos\theta\;}$$ > Pure-arithmetic dot product **equals** lengths times cosine of the angle. Arithmetic aur > geometry ek hi sikke ke do pehlu hain. **PICTURE.** Figure dono expressions ko stacked dikhata hai, $\|\mathbf u\|^2$ aur $\|\mathbf v\|^2$ dono sides par cross out hain, neeche boxed identity glowing hai. ![[deepdives/dd-maths-4.5.01-d2-s07.png]] Angle *dhundhne* ke liye rearrange kiya: $$\cos\theta = \frac{\mathbf u\cdot\mathbf v}{\|\mathbf u\|\,\|\mathbf v\|} \qquad\Longrightarrow\qquad \theta = \arccos\!\left(\frac{\mathbf u\cdot\mathbf v}{\|\mathbf u\|\,\|\mathbf v\|}\right).$$ Yahan $\arccos$ ("arc-cosine") is sawaal ka jawab deta hai *"kis angle ka yeh cosine hai?"* — yeh $\cos$ ko undo karta hai, bilkul waisi tarah jaise square root ek square ko undo karta hai. Hum iska output is page par **degrees mein** padhte hain (toh $\arccos(0)=90^\circ$). --- ## Step 8 — Har case: dot product ke sign ka kya matlab hai **KYA.** Lengths $\|\mathbf u\|,\|\mathbf v\|$ kabhi negative nahi hoti, toh $\mathbf u\cdot\mathbf v$ ka **sign** poori tarah $\cos\theta$ se decide hota hai. Aao *saare* cases walk karte hain taaki koi scenario surprise na kare. ($\mathbf u\cdot\mathbf v$ column mein har entry ek **plain number** hai, ek scalar — kabhi arrow nahi.) | $\theta$ | $\cos\theta$ | $\mathbf u\cdot\mathbf v$ | Geometry | |---|---|---|---| | $0^\circ$ | $+1$ | $+\|\mathbf u\|\|\mathbf v\|$ (max) | same direction | | $0^\circ$ aur $90^\circ$ ke beech | positive | positive | roughly aligned | | $90^\circ$ | $0$ | $0$ | **perpendicular** | | $90^\circ$ aur $180^\circ$ ke beech | negative | negative | roughly opposed | | $180^\circ$ | $-1$ | $-\|\mathbf u\|\|\mathbf v\|$ (min) | opposite directions | **PERPENDICULAR CASE SPECIAL KYUN HAI.** Exactly $\theta=90^\circ$ par, $\cos 90^\circ=0$, toh $\mathbf u\cdot\mathbf v=0$. Yeh workhorse test hai jo har jagah use hota hai [[Dot Product and Orthogonality]] mein: **zero dot product means right angle**, bina koi square roots ya angles kabhi compute kiye. **PICTURE.** Ek dial: jaise $\theta$ $0^\circ$ se $180^\circ$ tak sweep karta hai, arrow $\mathbf v$ ek fixed $\mathbf u$ se rotate hota jaata hai aur value $\mathbf u\cdot\mathbf v$ most-positive se, right angle par zero se hote hue, most-negative tak slide hoti hai. ![[deepdives/dd-maths-4.5.01-d2-s08.png]] > [!mistake] "Negative dot product zaroor ek error hai." > **Kyun galat lagta hai:** lengths positive hoti hain, toh surely product bhi positive hoga? > **Fix:** $\cos\theta$ factor negative ho jaata hai jab arrows $90^\circ$ se zyada apart point karte hain. > Negative dot product *meaningful* hai — yeh kehta hai "yeh somewhat opposite ways mein point karte hain." ### Degenerate case: agar ek arrow ki length zero ho toh? **KYA.** Agar $\mathbf u = \mathbf 0$ (zero arrow, ek single point bina kisi direction ke), tab $\mathbf u\cdot\mathbf v = 0$ automatically, aur $\|\mathbf u\| = 0$. **YEH SPECIAL CASE KYUN HAI.** Angle formula $\cos\theta = \dfrac{\mathbf u\cdot\mathbf v}{\|\mathbf u\|\|\mathbf v\|}$ $\|\mathbf u\|=0$ se divide karta hai — **undefined**. Ek point ki koi direction nahi hoti, toh "uski taraf angle" ek aisa sawaal hai jiska koi jawab nahi. Identity $\mathbf u\cdot\mathbf v=\|\mathbf u\|\|\mathbf v\|\cos\theta$ abhi bhi hold karta hai (dono sides $0$ hain), lekin aap $\theta$ ke liye *solve* nahi kar sakte. Divide karne se pehle hamesha zero vector check karo. --- ## Step 9 — Yahi proof $\mathbb{R}^n$ mein kyun kaam karta hai **KYA.** Steps 4–7 mein kabhi bhi coordinates nahi gine. Teen ingredients the: $\mathbf v\cdot\mathbf v = \|\mathbf v\|^2$ (Step 3), distributive/symmetric expansion (Step 6), aur Law of Cosines jo *triangle* par apply hua jise do arrows span karte hain (Step 5). **YEH KISI BHI DIMENSION MEIN KYUN LIFT KARTA HAI.** $\mathbb{R}^n$ mein dot product hai $\mathbf u\cdot\mathbf v = u_1v_1 + \cdots + u_nv_n$ — abhi bhi distributive aur symmetric, toh FOIL expansion word-for-word identical hai. Aur **koi bhi do arrows hamesha origin se ek single flat plane mein lie karte hain** (woh plane jo yeh span karte hain), toh Step 5 ka Law of Cosines triangle wahan bhi exist karta hai, chahe $n$ kitna bhi bada ho. Isliye: > [!formula] Result har $\mathbb{R}^n$ mein valid hai > $$\mathbf u\cdot\mathbf v = \|\mathbf u\|\,\|\mathbf v\|\cos\theta, > \qquad \mathbf u,\mathbf v\in\mathbb{R}^n.$$ > Do-dimensional picture scaffolding tha; algebra dimension-blind hai. --- ## Ek-picture summary ![[deepdives/dd-maths-4.5.01-d2-s09.png]] Poora derivation ek canvas par: $\mathbf u$, $\mathbf v$, $\mathbf u-\mathbf v$ ka triangle; teesra side **do tarike se** measure kiya gaya (geometry via Law of Cosines, algebra via dot-product FOIL); do expressions middle mein milte hain jahan $\|\mathbf u\|^2$ aur $\|\mathbf v\|^2$ cancel hote hain; aur boxed identity neeche drop out hoti hai. > [!recall]- Feynman: plain words mein poora walkthrough > Hamare paas do arrows the aur hum unke beech ka angle chahte the. Trick: unke tips connect karo triangle banane ke liye. > Ab humne connecting side ko **do alag-alag tarike se** measure kiya. Pehla tarika — geometry: Law of > Cosines us side ko do arrow-lengths aur angle se likhta hai. Doosra tarika — algebra: hum > difference arrow lete hain aur use dot product se khud se multiply karte hain, ise bilkul waisi tarah expand karte hain jaise > $(x-y)(x-y)=x^2-2xy+y^2$, aur middle term mein dot product $\mathbf u\cdot\mathbf v$ pop out karta hai. > Kyunki dono tarike *ek hi* side describe karte hain, hum unhe equal set karte hain. Bulky length-squared terms > dono sides par identical hain, toh yeh cancel ho jaate hain, aur jo bach jaata hai woh sirf yeh hai: dot product = length × > length × cosine of the angle. Magic yeh hai ki thodi si coordinate arithmetic — matching > numbers multiply karo, unhe add karo — secretly poora time ek angle measure kar rahi thi. Aur jab woh number > exactly zero hota hai, cosine zero hota hai, angle perfect right angle hai: isi tarah ek computer > perpendicularity *feel* karta hai bina kabhi picture draw kiye. Sabse achchi baat, humne ise plane mein draw kiya, lekin > arithmetic ne kabhi "kitne coordinates hain" nahi dekha, isliye exact same result > teen, chaar, ya sau dimensions mein bhi sach hai. > [!mnemonic] Proof paanch beats mein > **Triangle · Two ways · Set equal · Cancel · Cosine.** --- ## Active-recall Side $\mathbf u-\mathbf v$ kyun banate hain? ::: Yeh triangle ka teesra side hai, $\theta$ ke opposite, toh koi bhi angle-law isko involve karna chahiye. Kaunsa tool angle ko algebra mein laata hai, aur Pythagoras kyun nahi? ::: Law of Cosines; Pythagoras sirf $90^\circ$ handle karta hai, Law of Cosines koi bhi angle handle karta hai. Kya cheez humein $(\mathbf u-\mathbf v)\cdot(\mathbf u-\mathbf v)$ ko FOIL ki tarah expand karne deti hai? ::: Dot product distributive aur symmetric hai. Dono expressions ko equate karne par kya cancel hota hai? ::: $\|\mathbf u\|^2$ aur $\|\mathbf v\|^2$, dono sides par identically appearing. $\mathbf u\cdot\mathbf v=0$ ka geometrically kya matlab hai? ::: Arrows perpendicular hain ($\theta=90^\circ$). Zero vector ki taraf angle kyun nahi nikal sakte? ::: $\|\mathbf 0\|=0$, toh cosine formula zero se divide karta hai; ek point ki koi direction nahi hoti. 2-D proof $\mathbb{R}^n$ mein valid kyun rehta hai? ::: Algebra coordinates nahi ginती, aur koi bhi do arrows ek single plane span karte hain jahan Law of Cosines abhi bhi apply hota hai. --- ## Connections - [[Vectors in ℝⁿ — operations, geometric interpretation]] — parent note jise yeh walkthrough prove karta hai. - [[Dot Product and Orthogonality]] — Step 8 mein built zero-dot-product test use karta hai. - [[Norms and Distance in Rn]] — length $\|\mathbf v\|=\sqrt{\mathbf v\cdot\mathbf v}$ Step 3 se. - [[Projections and Orthogonal Decomposition]] — agli cheez jo $\cos\theta$ unlock karta hai.