Covariant and contravariant components: Visual walkthrough

4.10.8 · D2 · Maths › Advanced Topics (Elite Level) › Covariant and contravariant components

Step 1 — Ek blank tedha grid, aur ek seedha arrow

KYA. Do basis arrows draw karo jo usual square wale nahi hain: $e_{1}$ seedha right ki taraf point karta hai, $e_{2}$ ek slant pe upar-aur-right ki taraf point karta hai. Phir plane mein kahin ek single arrow $v$ rakho.

KYUN. Ordinary square graph paper pe jo do readings hum milne wale hain woh secretly identical hoti hain, isliye problem dikhti hi nahi. Do readings ko alag hote dekhne ke liye, hume paper ko jaanboojh kar tedha karna hoga. Ek basis bas un reference arrows ki jodi hai jinhe hum sab kuch measure karne ke liye choose karte hain.

PICTURE. Figure dekho. Pale grid lines do basis arrows ke parallel chalti hain — yahi tumhara tedha graph paper hai. Arrow $v$ (magenta) ek real geometric object hai: yeh exist karta hai pehle se, numbers choose karne se pehle.

Figure — Covariant and contravariant components

Step 2 — Reading #1: axes ke saath slide karo (contravariant)

KYA. Yeh poochho: main $e_{1}$ ke saath kitna chalta hoon, phir $e_{2}$ ke saath, taaki $v$ ki tip pe pahunch sakoon? Woh do travel-distances hi numbers $v^{1}$ aur $v^{2}$ hain.

KYUN. Yeh parallelogram rule hai — kisi bhi arrow ko do doosron se banane ka sabse natural tarika. Chhote upar ke numbers $v^{1}, v^{2}$ (index upar) contravariant components hain. Inhe is tarah define kiya jata hai: $v = steps along e_{1} v^{1} e_{1} + steps along e_{2} v^{2} e_{2} .$

PICTURE. Dashed lines ek parallelogram banati hain jiske sides axes ke parallel chalte hain. Orange path follow karo: $v^{1}$ $e_{1}$ ke saath chalo, mudo, $e_{2}$ ki copy ke saath $v^{2}$ chalo — tum exactly tip pe pahunch jaoge. Yahan woh path $1$ step phir $1$ step hai, isliye $v^{1} = 1, v^{2} = 1$ .

Step 3 — Reading #2: perpendicular giraya (covariant)

KYA. Ab parallelogram bhool jao. Iske badle, har axis ke liye, $v$ ka right-angle shadow seedha us axis pe daalo. Us shadow ki strength hai $v_{i} = v \cdot e_{i}$ .

KYUN. Dot product $v \cdot e_{i}$ literally $v$ ka $e_{i}$ ki direction mein perpendicular shadow measure karta hai (times $e_{i}$ ki length). Yeh Step 2 se alag sawaal ka jawaab deta hai: "main kitna slide karta hoon?" nahi balki " $v$ is axis se kitna agree karta hai?" Chhote neeche ke numbers $v_{1}, v_{2}$ (index neeche) covariant components hain: $v_{i} = humara arrow v \cdot axis i e_{i} .$

PICTURE. Violet dashed drop-lines har axis se right angle par milti hain (chhote square marks). Dhyan do ki shadow feet Step 2 ke parallelogram corners se alag jagahon pe land karti hain. Yahan $v_{1} = v \cdot e_{1} = 2$ aur $v_{2} = v \cdot e_{2} = 2 + 1 = 3$ .

Step 4 — Overlay: yeh disagree kyun karte hain

KYA. Dono readings ek hi picture pe rakho: parallelogram corners (Step 2) aur perpendicular feet (Step 3) ek saath.

KYUN. Inhe saath dekhne se poora subject click karta hai. Yeh sirf isliye differ karte hain kyunki axes $90°$ par nahi hain. Agar $e_{1} ⊥ e_{2}$ hota aur dono ki length $1$ hoti, toh parallelogram ek rectangle hota aur uske corners perpendicular feet ke exactly upar baithte — do readings merge ho jaati.

PICTURE. Orange dots (parallel) aur violet dots (perpendicular) coincide nahi karte. Woh gap dekho: yeh ek non-orthonormal basis ka visual signature hai. Inner product hi ek aisa tool hai jo us gap ko quantify kar sakta hai — yahi humara Step 5 hai.

Step 5 — Bridge dikhta hai: metric define karo

KYA. Ab hum do readings ke beech ka link derive karte hain. Covariant number ki definition se shuru karo aur contravariant expansion substitute karo: $v_{i} = v \cdot e_{i} = (v^{j} e_{j}) \cdot e_{i} = v^{j} sirf numbers! (e_{j} \cdot e_{i}) .$

KYUN. Covariant numbers $e_{i}$ ke saath dot karke define hote hain, aur humne pehle se $v = v^{j} e_{j}$ likha tha (Step 2) — toh hum bas ek ko doosre mein plug karte hain. Basis arrows ke dot products apne aap nikalte hain. Us table ko ek naam do: $g_{ij} \equiv e_{i} \cdot e_{j} ⟹ v_{i} = g_{ij} v^{j}$

$g_{ij}$ — metric tensor: ek chhoti symmetric table jo basis ki har length aur angle store karti hai.
$g_{ij} v^{j}$ — padho "har axis $i$ ke liye, contravariant numbers ko geometry table se mix karo" → covariant number nikalega.

PICTURE. Metric grid ka har cell ek overlap $e_{i} \cdot e_{j}$ dikhata hai. Diagonal cells $g_{11}, g_{22}$ squared lengths hain; off-diagonal $g_{12} = g_{21}$ slant measure karta hai. Humare basis ke liye: $g = (1112) .$

Step 6 — Return trip: inverse metric se raise karna

KYA. Covariant se contravariant wapas jaane ke liye, metric ko uske inverse $g^{ij}$ se undo karo: $v^{i} = g^{ij} v_{j}, g^{ij} g_{j k} = δ_{k}^{i} .$

KYUN. $g_{ij}$ ek matrix hai; matrix multiply ko reverse karne ka ek hi sahi tarika hai uske inverse se multiply karna. Symbol $δ_{k}^{i}$ (Kronecker delta, $1$ agar $i = k$ warna $0$ ) identity table hai — " $g$ phir $g^{- 1}$ exactly wahi wapas deta hai jo tumne shuru kiya tha."

PICTURE. Figure derivation ke do arrows dikhata hai: $g_{ij}$ staircase se neeche point karta hai (raise→lower), $g^{ij}$ wapas upar. Humare basis ke liye $g^{- 1} = (2 - 1 - 1 1)$ , aur $g^{ij} v_{j}$ wapas $(2, 3) \to (1, 1)$ return karta hai.

Step 7 — Degenerate case: square paper (kyun school ne kabhi nahi bataya)

KYA. Sab kuch $e_{1} ⊥ e_{2}$ ke saath redraw karo, dono length $1$ — ek orthonormal basis.

KYUN. Yeh sanity check hai jise har claim ko survive karna chahiye. Ab $e_{i} \cdot e_{j} = δ_{ij}$ hai, isliye metric identity mein collapse ho jaata hai, aur $v_{i} = g_{ij} v^{j} = δ_{ij} v^{j} = v^{i}$ . Do readings ek mein fuse ho jaati hain.

PICTURE. Parallelogram ek rectangle ban jaata hai; uske corners perpendicular feet ke upar land karte hain. Orange aur violet dots coincide hote hain — Step 4 ka gap gayab ho gaya. Yahi wajah hai ki school mein ordinary vectors ke liye sirf ek set of components ki zaroorat hoti hai.

Recall Edge cases ek nazar mein

Square paper metric ko identity mein collapse karta hai? ::: Haan — $g_{ij} = δ_{ij}$ , isliye $v_{i} = v^{i}$ . Agar do axes parallel hon (same direction)? ::: Basis fail ho jaata hai: koi parallelogram close nahi hota, $g$ singular hai (koi inverse nahi), tum indices raise nahi kar sakte. Longer axis $e_{2}$ lekin phir bhi perpendicular? ::: $g$ diagonal rehta hai lekin identity nahi; readings per axis ek scale factor se differ karti hain, koi cross-mixing nahi.

Ek-picture summary

Sab kuch ek canvas pe: arrow $v$ , uski do readings, aur metric bridge jo unhe jodhta hai. Loop trace karo — parallel slide se $v^{i}$ milta hai, perpendicular shadow se $v_{i}$ milta hai, aur $g_{ij}$ / $g^{ij}$ tumhe unke beech le jaate hain. Yahi exact machinery curvilinear coordinates mein aur general relativity mein dobara dikhti hai, jahan metric constant nahi hota.

Recall Poore walkthrough ki Feynman retelling

Maine tedha graph paper draw kiya aur uspe ek arrow rakha. Maine do fair sawaal poochhe. Sawaal ek: main har tedhi line ke saath arrow ki tip tak pahunchne ke liye kitna chalta hoon? — isse mujhe "up-index" numbers mile, contravariant wale $(1, 1)$ . Sawaal do: arrow ka seedha neeche gira shadow har line pe kitna bada hai? — isse mujhe "down-index" numbers mile, covariant wale $(2, 3)$ . Tedhe paper pe yeh do jawab disagree karte hain, aur yeh disagreement bilkul sirf isliye hoti hai ki axes ninety degrees par nahi hain. Toh maine ek chhoti translation table banayi basis arrows ko ek doosre ke saath dot karke — metric $g$ . $g$ se multiply karna up-numbers ko down-numbers mein badalta hai; uske inverse se multiply karna unhe wapas le aata hai. Aakhir mein maine paper ko squares mein seedha kiya: parallelogram ek rectangle ban gaya, do jawab snap karke saath aa gaye, aur poora mystery gayab ho gaya — yahi exactly wajah hai ki school mein kisi ne kabhi nahi bataya ki components do tarah ke hote hain.