4.10.11 · D2 · HinglishAdvanced Topics (Elite Level)
Visual walkthrough — Christoffel symbols — intro
4.10.11 · D2· Maths › Advanced Topics (Elite Level) › Christoffel symbols — intro
Neeche jo bhi hai woh teen prerequisite pictures par tika hai. Agar koi bhi word naya lage, link follow karo: Metric Tensor, Polar Coordinates, Covariant Derivative, Tensor Transformation Laws.
Step 0 — Woh do pictures jinpar har symbol rehta hai
Kisi bhi formula se pehle, do mental images fix karo.
KYA HAI. Ek basis vector ek chota sa arrow hai jo kehta hai "coordinate number is direction mein badhta hai." Flat Cartesian duniya mein do aise arrows hote hain, aur , aur woh har point par same arrow hote hain. Polar duniya mein (bahar ki taraf point karta hai) aur (sideways point karta hai, circle ke saath saath) hote hain — aur yeh mudtey hain jab tum move karte ho.
KYUN. Poora subject isliye exist karta hai kyunki woh arrows mudtey hain. Agar woh kabhi nahi mudtey, toh kuch bhi bookkeep karna nahi padta aur har jagah zero hota.
PICTURE. Left par, Cartesian arrows: identical copies. Right par, ek circle par teen points par polar arrows — dekho kaise lal ghoomta hai.
Hum metric use karte hain aur raw arrows nahi, kyunki ek real problem mein (ek warped surface) metric woh ek cheez hai jo hum actually measure kar sakte hain — dekho Metric Tensor.
Step 1 — Us twist ko naam do jo hum measure karna chahte hain
KYA HAI. Jab tum direction mein thoda nudge karte ho, arrow thoda naya arrow ban jaata hai. Woh change khud ek arrow hai. Koi bhi arrow original basis arrows ki recipe ke roop mein rebuild kiya ja sakta hai:
=\;\underbrace{\Gamma^{k}_{\;ij}}_{\substack{\text{us change mein }\mathbf e_k\\ \text{ki matra}}}\;\underbrace{\mathbf{e}_k}_{\text{basis arrow }k}$$ Equation ko term by term zor se padho: "arrow $i$ mein change jab main direction $j$ mein step karta hoon (**left**) barabar hai original arrows ka ek dher (**right**), jahan $\Gamma^{k}_{\;ij}$ us dher mein arrow $k$ ka *serving size* hai." **Aise define kyun karein?** Kyunki hum change ko **usi language mein express karna chahte hain jo hamare paas pehle se hai** — original basis. Yahi ek tarika hai jisse ek computer (ya ek covariant derivative) use kar sake. **PICTURE.** Ek green "change arrow" ko ek lal $\mathbf{e}_r$-portion aur ek neele $\mathbf{e}_\theta$-portion mein decompose kiya gaya hai. Woh do portions hi $\Gamma^{r}_{\;ij}$ aur $\Gamma^{\theta}_{\;ij}$ hain. ![[deepdives/dd-maths-4.10.11-d2-s02.png]] > [!recall]- Kyun do lower indices symmetric hote hain > Lower indices ki symmetry ::: Kyunki $\partial_j\partial_i(\text{position})=\partial_i\partial_j(\text{position})$ — mixed partials commute karte hain. Kyunki $\mathbf e_i=\partial_i(\text{position})$, hum paate hain $\partial_j\mathbf e_i=\partial_i\mathbf e_j$, isliye $\Gamma^k_{\;ij}=\Gamma^k_{\;ji}$. --- ## Step 2 — Problem: $\Gamma$ dot products ke *andar* chhupa hai, khule mein nahi Hum $g$ se $\Gamma$ chahte hain, lekin $\Gamma$ arrows ke derivatives mein rehta hai aur $g$ arrows ke dot products mein rehta hai. Bridge ek **beech ka object** hai: change-arrow ko ek plain basis arrow se dot karo. **KYA HAI.** **First-kind** symbol define karo $$\Gamma_{ijk}=(\partial_j\mathbf e_i)\cdot\mathbf e_k .$$ Term by term: $j$ ke saath $\mathbf e_i$ ka change lo (**ek arrow**), use $\mathbf e_k$ se dot karo (**use arrow $k$ par project karo**). Result ek plain number hai — shadow ki length. **KYUN.** Dotting arrows ko numbers mein badal deta hai, aur numbers hi metric ki material hai. Yahi poori derivation mein ek trick hai: *twist ko ek jaane-pehchaane arrow par project karo.* **PICTURE.** Green change-arrow $\mathbf e_k$ par shadow daalta hai; shadow ki length $\Gamma_{ijk}$ hai. ![[deepdives/dd-maths-4.10.11-d2-s03.png]] Second kind se relation (index ko inverse metric $g^{km}$ se raise karna, jo ek projection-length ko wapas serving-size mein convert karta hai): $$\Gamma^{k}_{\;ij}=g^{km}\,\Gamma_{ijm}.$$ --- ## Step 3 — Metric ko differentiate karo (woh product rule jo sab kuch shuru karti hai) **KYA HAI.** Metric hai $g_{ij}=\mathbf e_i\cdot\mathbf e_j$. Dot products ke liye product rule use karke $x^k$ ke saath differentiate karo: $$\partial_k g_{ij} =\underbrace{(\partial_k\mathbf e_i)\cdot\mathbf e_j}_{\Gamma_{ikj}} +\underbrace{\mathbf e_i\cdot(\partial_k\mathbf e_j)}_{\Gamma_{jki}}.$$ Har half Step 2 se exactly ek first-kind symbol hai. Toh: $$\boxed{\;\partial_k g_{ij}=\Gamma_{ikj}+\Gamma_{jki}\;}\tag{A}$$ **KYUN.** Yeh *ek hi* fact hai jo "metric ke derivatives" (cheezein jo hum compute kar sakte hain) ko "$\Gamma$'s" (cheezein jo hum chahte hain) se link karta hai. Ek equation, do unknown $\Gamma$'s — abhi solve nahi hoga. Yahi Step 4 ko motivate karta hai. **PICTURE.** Dot product mein change = "$\mathbf e_i$ ka twist $\mathbf e_j$ par shadow" plus "$\mathbf e_j$ ka twist $\mathbf e_i$ par shadow." ![[deepdives/dd-maths-4.10.11-d2-s04.png]] --- ## Step 4 — Teen copies, cycle kiye hue (woh trick jo ek $\Gamma$ ko isolate karti hai) **KYA HAI.** Equation (A) **kisi bhi** teen slots ki choice ke liye hold karta hai. Ise teen baar likho, har baar labels $i\to j\to k\to i$ rotate karte hue: $$\partial_k g_{ij}=\Gamma_{ikj}+\Gamma_{jki}\tag{1}$$ $$\partial_i g_{jk}=\Gamma_{jik}+\Gamma_{kij}\tag{2}$$ $$\partial_j g_{ki}=\Gamma_{kji}+\Gamma_{ikj}\tag{3}$$ **Teen kyun?** Hamare paas teen unknown first-kind symbols aapas mein uljhe hue hain. Teen equations bilkul enough hain ek ke liye solve karne ko. Cyclic rotation guarantee karta hai ki har $\Gamma$ ek *predictable* jagah par aata hai taaki Step 5 mein cancellation kaam kare. **PICTURE.** Ek wheel jis par $i,j,k$ hain; wheel ka har chakkar teen equations mein se ek produce karta hai. ![[deepdives/dd-maths-4.10.11-d2-s05.png]] Step 1 se lower-index symmetry yaad karo, first kind ke liye likhi gayi: $\Gamma_{abc}=\Gamma_{bac}$ (pehle do slots ko freely swap karo — woh do coordinate directions hain jinका order matter nahi karta). --- ## Step 5 — Do jodo, ek ghataao, dekho chaar terms gayab ho jaate hain **KYA HAI.** (2) $+$ (3) $-$ (1) compute karo. Right-hand side ke chhe $\Gamma$'s ko line up karo: $$\big(\Gamma_{jik}+\Gamma_{kij}\big)+\big(\Gamma_{kji}+\Gamma_{ikj}\big)-\big(\Gamma_{ikj}+\Gamma_{jki}\big).$$ Ab $\Gamma_{abc}=\Gamma_{bac}$ use karo: - $+\Gamma_{ikj}$ (3 se) aur $-\Gamma_{ikj}$ (1 se) **seedha cancel ho jaate hain**. - $\Gamma_{jik}=\Gamma_{ijk}$ aur $\Gamma_{jki}=\Gamma_{jki}$: pair $+\Gamma_{jik}-\Gamma_{jki}$... symmetry apply karo $\Gamma_{jki}=\Gamma_{kji}$, aur $+\Gamma_{jik}$ remaining $+\Gamma_{kij}$ aur $+\Gamma_{kji}$ ke saath pair karta hai taaki **do surviving terms equal ho jaate hain**, giving $2\Gamma_{kij}$. Sabse saaf bookkeeping: sab kuch collapse ho jaata hai $$\partial_i g_{jk}+\partial_j g_{ki}-\partial_k g_{ij}=2\,\Gamma_{kij}.$$ 2 se divide karo — *wahan hai tumhara $\tfrac12$*: $$\boxed{\;\Gamma_{kij}=\tfrac12\big(\partial_i g_{jk}+\partial_j g_{ki}-\partial_k g_{ij}\big)\;}\tag{B}$$ **Two-plus-one-minus pattern kyun?** Woh do equations jo humne *jode* (2 aur 3) dono mein woh $\Gamma$ tha jo hum chahte the, toh woh reinforce hota hai (isliye double hota hai → $\tfrac12$ doubling undo karta hai). Jise humne *ghataaya* (1) woh sirf leftover $\Gamma_{ikj}$ junk ko kill karne ke liye tha. **PICTURE.** Chhe tiles: do green glow karti hain (rakhi gayi, $2\Gamma$ ban jaati hain), do grey glow karti hain (cancel ho jaati hain), do merge ho jaati hain. ![[deepdives/dd-maths-4.10.11-d2-s06.png]] --- ## Step 6 — Index raise karo: shadow-length se serving-size tak **KYA HAI.** Equation (B) *first-kind* $\Gamma_{kij}$ deta hai — ek projection length. Serving-size $\Gamma^{k}_{\;ij}$ paane ke liye hum **inverse metric** $g^{km}$ se contract karte hain (woh table jo $g$ ko undo karti hai, yaani $g^{km}g_{mj}=\delta^k_{\;j}$): $$\Gamma^{k}_{\;ij}=g^{km}\,\Gamma_{mij} =\tfrac12\,g^{km}\big(\partial_i g_{mj}+\partial_j g_{mi}-\partial_m g_{ij}\big).$$ Ab finished formula ko un sab cheezein ke saath padho jo humne banayeen: - $\tfrac12$ — Step 5 mein doubling se. - $g^{km}$ — inverse metric jo ek shadow ko arrow $k$ ki serving mein badle (Step 6). Iska dummy index $m$ woh hai jo **minus** carry karta hai. - $+\partial_i g_{mj}+\partial_j g_{mi}$ — woh do *jodi gayi* equations (Step 4/5). - $-\partial_m g_{ij}$ — woh ek *ghatayi gayi* equation, dummy index par. ![[deepdives/dd-maths-4.10.11-d2-s07.png]] > [!formula] Result, poori tarah se kamaaya hua > $$\boxed{\;\Gamma^{k}_{\;ij}=\tfrac12\,g^{km}\big(\partial_i g_{mj}+\partial_j g_{mi}-\partial_m g_{ij}\big)\;}$$ --- ## Step 7 — Flat degenerate case: sab kuch zero ho jaana chahiye **KYA HAI.** Cartesian coordinates mein arrows kabhi nahi mudtey, toh $g_{ij}=\delta_{ij}$ (diagonal par 1, off it 0) — **constant** har jagah. Har derivative $\partial_m g_{ij}=0$, toh har bracket $0+0-0=0$ hai, isliye **har $\Gamma=0$**. **KYUN DIKHAAYEN?** Ek formula jo un arrows ke liye nonzero twist deta jo kabhi twist nahi karte, woh galat hota. Yeh woh sanity check hai jo formula ko paas karna chahiye — aur karta hai, kyunki yeh sirf $g$ ke derivatives se bana hai. **PICTURE.** Cartesian arrows: change-arrow (Step 2 ka green) ek single dot mein sir jaata hai. Project karne ko kuch nahi, serve karne ko kuch nahi. ![[deepdives/dd-maths-4.10.11-d2-s08.png]] > [!mistake] "Flat matlab sab $\Gamma$ vanish." > **Kyun sahi lagta hai:** upar ka Cartesian case. > **Fix yeh hai:** flat *polar* coordinates mein $\Gamma^{r}_{\;\theta\theta}=-r\neq0$ milta hai. Arrows mudtey hain bhi jab space flat ho. $\Gamma$ ka vanish hona metric ke derivatives ka zero hona require karta hai, yaani ek *coordinate* fact, curvature fact nahi. Curvature [[Riemann Curvature Tensor]] mein rehti hai jo $\partial\Gamma+\Gamma\Gamma$ se bana hai. --- ## Step 8 — Numbers mein non-degenerate case: polar coordinates **KYA HAI.** Polar metric $g_{rr}=1,\ g_{\theta\theta}=r^2,\ g_{r\theta}=0$, inverse $g^{rr}=1,\ g^{\theta\theta}=1/r^2$. Boxed formula mein feed karo. $$\Gamma^{r}_{\;\theta\theta}=\tfrac12 g^{rr}\big(\underbrace{\partial_\theta g_{r\theta}}_{0}+\underbrace{\partial_\theta g_{r\theta}}_{0}-\underbrace{\partial_r g_{\theta\theta}}_{2r}\big)=\tfrac12(1)(-2r)=\boxed{-r}$$ $$\Gamma^{\theta}_{\;r\theta}=\tfrac12 g^{\theta\theta}\big(\underbrace{\partial_r g_{\theta\theta}}_{2r}+0-0\big)=\tfrac12\cdot\tfrac1{r^2}\cdot 2r=\boxed{\tfrac1r}$$ **KYUN.** Yeh Step 0 ke right panel mein jo twist humne *dekha* tha usi se match karte hain: bahar jaate waqt ($r$ badhta hai), $\theta$-arrow stretch karta hai, $1/r$ deta hai; ghoomte waqt, $\theta$-arrow ki tip andar curl karti hai, $-r$ deta hai. Pictures aur algebra agree karte hain. **PICTURE.** Polar arrow-field jis par do surviving $\Gamma$'s exactly wahin label hain jahaan unka twist hota hai. ![[deepdives/dd-maths-4.10.11-d2-s09.png]] Yahi numbers [[Geodesic Equation]] ($\ddot r - r\dot\theta^2=0$) ko drive karte hain — $-r$ centripetal term hai. Aur yeh [[Covariant Derivative]] $\nabla_j V^k=\partial_j V^k+\Gamma^{k}_{\;ij}V^i$ mein fit ho jaate hain. --- ## The one-picture summary ![[deepdives/dd-maths-4.10.11-d2-s10.png]] Ek nazar mein, poora raasta: arrows mudtey hain → twist ko project karo ($\Gamma_{ijk}$) → metric differentiate karo → teen cyclic copies → do jodo ek ghataao → 2 se divide karo → $g^{km}$ se raise karo. > [!recall]- Feynman: poora walkthrough plain words mein > Retelling ::: Tumhare paas direction-arrows hain jo chalte waqt mudtey hain. Tum chahte ho ek number ki kitna mudta hai. Tum seedha twist nahi dekh sakte, toh tum iska **shadow** ek neighbouring arrow par pakad lete ho — woh shadow ek "first-kind" number hai. Shadows wahi stuff hain jisse dot products bane hain, aur dot products hi metric *hain*. Toh tum metric differentiate karte ho aur, sure enough, woh do shadows ka sum nikalta hai. Ek equation do shadows untangle karne ke liye enough nahi hai, toh tum labels ghuma ke teen baar same equation likhte ho. Unme se do jodo aur ek ghataao: junk cancel ho jaata hai, joh shadow tum chahte the woh double ho jaata hai, toh tum do se divide karte ho. Aakhir mein shadow-length ko ek honest "arrow $k$ kitna" serving mein badale ke liye inverse metric $g^{km}$ se multiply karte ho — aur boxed formula nikal aata hai, minus sign exactly dummy index par baithe hue, halves aur pluses sab account kiye hue. Ise flat Cartesian arrows ke against check karo (sab zero — good) aur polar arrows ke against (gives $-r$ aur $1/r$, centripetal aur Coriolis terms jo tum literally arrows ko karte hue dekh sakte ho). --- ## Active-Recall Change-arrow ko ek basis arrow se dot karna kyun help karta hai? ::: Yeh ek arrow ko number mein badal deta hai, aur numbers hi metric (dot products) ki material hai — $\Gamma$ se $g$ tak bridge. Step 5 mein $\tfrac12$ kahan se aata hai? ::: Joh $\Gamma$ chahiye tha woh dono jodi gayi equations mein appear hota hai, isliye double ho jaata hai; 2 se divide karna doubling undo karta hai. Final formula mein minus sign kaun sa term carry karta hai? ::: $-\partial_m g_{ij}$, dummy index $m$ ke respect mein derivative jo $g^{km}$ se tied hai. Cartesian coordinates mein sab $\Gamma$ kyun vanish karte hain? ::: $g_{ij}=\delta_{ij}$ constant hai, toh har $\partial_m g_{ij}=0$; bracket zero hai. Last step mein $g^{km}$ kya role play karta hai? ::: Yeh index raise karta hai, ek projection-length (first kind) ko arrow $k$ ki serving-size (second kind) mein convert karta hai. --- ## Connections - [[Christoffel symbols — intro]] — woh parent jiska boxed formula humne zero se grow kiya. - [[Metric Tensor]] — dot-product table jis par har step depend karta hai. - [[Polar Coordinates]] — kaam kiya hua non-degenerate case. - [[Covariant Derivative]] — jahaan yeh $\Gamma$'s use hotey hain. - [[Geodesic Equation]] — woh physics jo $-r$ ko centripetal padhti hai. - [[Riemann Curvature Tensor]] — $\partial\Gamma+\Gamma\Gamma$ se bana, asli curvature. - [[Tensor Transformation Laws]] — kyun $\Gamma$ ek tensor *nahi* hai.