how do we know the derivative of ln(x) is 1/x (the definition & implicit differentiation)

Поделиться
HTML-код
  • Опубликовано: 14 дек 2024

Комментарии • 672

  • @wildramen
    @wildramen 7 лет назад +882

    I have graduated 3 months ago, at the start of the calculus class 2 years ago i hated calculus but here i am, loving calculus and enjoying every second of your awesome videos.

    • @blackpenredpen
      @blackpenredpen  7 лет назад +75

      Alex Ramyeon thank you!!!!!!

    • @tintinfan007
      @tintinfan007 2 года назад +8

      Honestly speaking calculus is more fun than GTA and all other video games.

    • @ct---cp8li
      @ct---cp8li Год назад

      @@blackpenredpen my way docs.google.com/document/d/e/2PACX-1vQ0SB1cs5gR0S17zmIhfFuQqmhGsw8_jn_QoL1n6AjI26wsu2bOPIxzCrw1D0SK-fCca1FUR-xAQ-gI/pub

    • @aaxel_dz
      @aaxel_dz Год назад +2

      @@tintinfan007 you're probably speaking about some random mobile gta spinoff

    • @climbeverest
      @climbeverest Год назад +2

      I love the way he teaches

  • @idomins
    @idomins 6 лет назад +1526

    You know its about to get real when he starts using the blue pen (-:

    • @blackpenredpen
      @blackpenredpen  6 лет назад +157

      yup, that's right!

    • @ciherrera
      @ciherrera 6 лет назад +6

      Somewhat relevant xkcd: xkcd.com/849/

    • @Peter_1986
      @Peter_1986 4 года назад +28

      He has also used a purple pen a few times.

    • @leviackerman6090
      @leviackerman6090 4 года назад +2

      😂😂😂😂😂😂😂😂

    • @in-ty8vb
      @in-ty8vb 3 года назад +2

      Multiple Colour Pen

  • @yoavshati
    @yoavshati 7 лет назад +1724

    8:11
    Talking to your GF

  • @soyalguien335yt4
    @soyalguien335yt4 2 года назад +166

    10:43 I really love the satisfaction I get when my mind snaps and know how the demostration continues before the video. Great video!

    • @vvvss-cx1vd
      @vvvss-cx1vd 2 года назад +2

      Clicked at around 7:50 for me, so satisfying

    • @BradleyG01
      @BradleyG01 2 года назад +2

      was about to cmment that same thing! It's such a great feeling

  • @whozz
    @whozz 7 лет назад +326

    Another proof using parametric equation:
    x = e^t
    y = t
    dx/dt = e^t
    dy/dt = 1
    (dy/dt)/(dx/dt) = dy/dx = 1/e^t = 1/x

    • @rajendramisir3530
      @rajendramisir3530 6 лет назад +42

      Wonder proof buddy! Three different proofs: Limits, implicit differentiation and parametric equations.

    • @rajendramisir3530
      @rajendramisir3530 6 лет назад +16

      Wonderful proof buddy! Three different proofs: Using parametric equations, limits and implicit differentiation.

    • @DaveJ6515
      @DaveJ6515 4 года назад +34

      Just brilliant. Congratulations, I’m going to teach this one tomorrow

    • @pradipgiri8321
      @pradipgiri8321 4 года назад +7

      nicely done

    • @RamsLiff
      @RamsLiff 4 года назад +4

      For any log , 1/x .ln a, a its the base of the log , If a = e, the derivative is 1/x
      I did a general proof

  • @jimbeasley5319
    @jimbeasley5319 5 лет назад +356

    How do we know ln(x) is a logarithm? I once had a professor “define” ln(x) as a function whose derivative is 1/x. He then proceeded to show the ln(x) is indeed a logarithm, and it has the base e. I’d like to see this again. It was very inspiring, but I have forgotten how it was done.

    • @blackpenredpen
      @blackpenredpen  5 лет назад +108

      Are you asking how to prove the properties based on that definition? If so, I have a video here ruclips.net/video/4D-M5qB_l6k/видео.html

    • @tckgkljgfl7958
      @tckgkljgfl7958 4 года назад +2

      @hobo doc id be happy to receive those pages on scrubster@gmx.de

    • @lewisbotterill4948
      @lewisbotterill4948 3 года назад +84

      I was taught that ln(x) is by definition, log base e of x. The term ln itself means the natural logarithm.

    • @cellcomsanggau424
      @cellcomsanggau424 3 года назад +6

      @@lewisbotterill4948 look

    • @lewisbotterill4948
      @lewisbotterill4948 3 года назад +18

      @@cellcomsanggau424 ?

  • @muhammedyaseen5438
    @muhammedyaseen5438 4 года назад +221

    8:25 my brain to me after a test

  • @sajidrafique375
    @sajidrafique375 4 месяца назад +2

    I loved calculus in college and now i am 71 and watch these videos to fill the gaps in my understanding and keep dementia away

  • @ezvac1421
    @ezvac1421 3 года назад +52

    Another way :
    exp(ln(x)) = x
    Derivative of both sides :
    ln(x)' * exp(ln(x)) = 1
    Replace exp(ln(x)) by x and divide the whole equation by it :
    ln(x)' = 1/x

  • @magellan500
    @magellan500 2 года назад +20

    Just watched it again as there were a few things I wasn’t sure of. I really liked it when he explained one trick to use was because the natural log is a continuous function, and the limit of a continuous function is a continuous function of the limit, you can move the limit inside the parentheses to simplify things. Cool stuff.

  • @itamarbentwich3814
    @itamarbentwich3814 7 лет назад +9

    I really enjoyed your last few videos, and I am glad you're back to uploading more videos containing your explanations

  • @taranmellacheruvu2504
    @taranmellacheruvu2504 2 года назад +6

    You can also use the formula for inverse derivatives. This is how I did it:
    Let g(x) = the inverse of f(x)
    g’(x) = 1/(f’(g(x))
    Let f(x) = e^x
    Therefore f’(x) = e^x & g(x) = lnx
    g’(x) = 1/(e^lnx)
    g’(x) = 1/x
    Therefore the derivative of lnx is 1/x.
    To prove the formula I used, you can let
    g(x) = inverse of f(x)
    So,
    x = f(g(x))
    Differentiating both sides, you get:
    1 = f’(g(x))*g’(x)
    g’(x) = 1/(f’(g(x))

    • @xnqmap
      @xnqmap 2 года назад

      It basically what he does from 13:00, without explicitly using the formula for the derivative of a reciprocal function.

  • @FF-pv7ht
    @FF-pv7ht 7 лет назад +280

    Its a shame we dont get teached this stuff in school but are just supposed to remember f'(x)=1/x of F(x)=ln(x)

    • @ChaosPod
      @ChaosPod 7 лет назад +40

      I remember my school teaching us a variation of the 2nd method, namely
      y = ln x
      => e^y = x
      Therefore dx/dy = e^y
      dy/dx = 1/(dx/dy) = 1/e^y = 1/x

    • @Witiok1992
      @Witiok1992 6 лет назад

      FF same situation(((

    • @ZZaarraakkii
      @ZZaarraakkii 6 лет назад +7

      Of f(x)=ln(x). Capital f often implies integration. Especially because integrated function is defined by it F'(x)=f(x) then you are ok.

    • @sjoerdo6988
      @sjoerdo6988 6 лет назад +22

      they told us:
      e^ln(x)=x
      diferentiating gives:
      e^ln(x)*d/dx(ln(x))=1
      d/dx(ln(x))=1/e^ln(x)=1/x

    • @znhait
      @znhait 6 лет назад +8

      This is just the application of the first principle definition of the derivative. You know how to do limits and should be well versed in algebra manipulation. It's not a big leap to do this problem. This is the sort of exercise a student should do away from school.

  • @nobels3246
    @nobels3246 Год назад +1

    To differentiate ln(x) I use this trick:
    1 = 1
    1 = d/dx x
    1 = d/dx [e^(ln(x))]
    1 = e^(ln(x)) * d/dx(ln(x))
    d/dx(ln(x)) = 1 / [e^(ln(x))]
    d/dx(ln(x)) = 1/x
    This also works for all inverse functions like arcsin(x), arcos(x) & arctan(x).

  • @iansheridan5509
    @iansheridan5509 4 года назад +3

    You are my new favorite high school math teacher. In my AP calculus class, we were never taught how to derive this. Only taught to memorize that d/dx ln(x) = 1/x

  • @gebcrafter
    @gebcrafter 5 лет назад +96

    this math professor dripping out with tha preme jacket

  • @hopp2184
    @hopp2184 7 лет назад +69

    blackpenredpen could you solve the non elementary integral of x^x. You did the (easier) derivative so please do the difficult integral or let Payem do it

    • @angelmendez-rivera351
      @angelmendez-rivera351 5 лет назад +3

      Ahsoka Tano How is he suppose to solve it if it is non-elementary? Do you understand what solving an integral is? And do you understand what non-elementary is?

    • @zachcate7102
      @zachcate7102 5 лет назад +13

      Angel Mendez-Rivera ima be real with you that made no sense

    • @nicememe8608
      @nicememe8608 5 лет назад +5

      Zach Cate if an integral is non-elementary, by definition, that means you cannot solve it. It will be defined by a special function. For example, the fresnel integral

  • @marcushendriksen8415
    @marcushendriksen8415 6 лет назад +35

    Great video man! I feel like you've made me so much smarter; this time I was actually able to see ahead a little bit, that the argument of ln would end up being e^1/x (this was around when you brought the derivative into the u world)

  • @AkiyamaKatsuko
    @AkiyamaKatsuko Год назад +2

    I've always been told that the derivative rule for f'(x) of ln(x) has always been 1/x but I've never understood how that was proven. Thank you for the explanation.

    • @carultch
      @carultch Год назад

      There's usually one of these proofs for it somewhere in the textbook. Since the teacher probably sees proving them as reinventing the wheel, and not necessary to understand the subject, they probably just skip showing why these derivative rules work.

  • @TheOriginalJohnDoe
    @TheOriginalJohnDoe Год назад

    12:05 when I saw this, I was like... OMG I just realized what the hell I've been watching for the past 12 minutes... I was more intrigued by what he was able to do in terms of modifying the formulae, but then noticed he brought it down to 1/x, I love this guy.

  • @rajendramisir3530
    @rajendramisir3530 6 лет назад +3

    Elegant proofs for the derivative of ln(x). I like the intelligent and creative ways you used to develop and establish your proofs. Thanks.

  • @FurryCombatWombat
    @FurryCombatWombat 3 года назад

    This mad lad really just used the limit definition. Can we get this guy a medal?

  • @KD_elctrcL_N_elctrnX
    @KD_elctrcL_N_elctrnX 2 года назад +3

    How easily he changes markers is amazing to watch

  • @jakehu8733
    @jakehu8733 11 месяцев назад +1

    I have a fourth proof: If we differentiate e^ln x, instead of resulting in x, we use the chen lu, where u = ln(x). That results in e^(ln x) * du/dx. However, if we use the power rule, it results in 1. Therefore, x * du/dx = 1. We solve for du/dx = 1/x.

  • @mastershooter64
    @mastershooter64 4 года назад +5

    dayuumm now that's impressive, finding the derivative of ln(x) using the standard definition of a derivative

  • @magellan500
    @magellan500 2 года назад +3

    I have always wanted a more detailed explanation of this result. This is the best I’ve seen on the subject. Considering things like Euler’s identity and the quantum wave equation and other uses of the exponential function, it seems to me it’s the most useful of all the special functions.

  • @brianbarringer3458
    @brianbarringer3458 8 месяцев назад +1

    Thanks!

  • @SudarshanBaurai
    @SudarshanBaurai Год назад

    Dear friend, you are not only genius but you a great guru (teacher). My regards - Sudarshan🙏

  • @papajack2205
    @papajack2205 7 лет назад +51

    Finally, since the basketball secret has been revealed I can find some sleep!

    • @blackpenredpen
      @blackpenredpen  7 лет назад +4

      Marian P. Gajda in fact, in was in that previous video as well, just allllll the way at the end.

    • @derekanderson1214
      @derekanderson1214 7 лет назад +2

      I live close to where you recorded that basketball video! I was pleasantly surprised when I saw that

    • @blackpenredpen
      @blackpenredpen  7 лет назад +7

      Derek Anderson
      Are you serious???
      How did u even recognize that place!!!

    • @aashsyed1277
      @aashsyed1277 3 года назад

      @@blackpenredpen COINCIDENCE?

    • @aashsyed1277
      @aashsyed1277 3 года назад

      @@blackpenredpen it is possible, but very unlikely

  • @FocusLRHAP
    @FocusLRHAP 6 месяцев назад +1

    16:43 Yeah! It even works for e: ln(e)=1 so you get back the 1/x :)

  • @hypereric5526
    @hypereric5526 3 года назад

    I love these type of people on RUclips

  • @lotaniq4449
    @lotaniq4449 2 года назад +2

    y=lnx
    e^y=x
    (e^y)’=x’
    (e^y)*y’=1
    y’=1/(e^y)
    (lnx)’=1/x (substitution of the given parts)

    • @AhmedKhan-qk3xi
      @AhmedKhan-qk3xi 2 года назад

      The proof of the derivative of e^x uses the exact same ln lim n->0 so it’s better to derive with limits and not other operations that rely on said limits because it can end up begging the question

    • @lotaniq4449
      @lotaniq4449 2 года назад

      @@AhmedKhan-qk3xi idk what u are saying but ig ur right

    • @lotaniq4449
      @lotaniq4449 2 года назад

      @@AhmedKhan-qk3xi I think it is just way shorter

  • @philippenachtergal6077
    @philippenachtergal6077 2 года назад +1

    6:30 Hum. Ah. But when we say limit(h->0) that implies in any direction right ?
    As we work with real numbers we can have h < 0 or h > 0 and both directional limits (or whatever the proper name for that is) must give the same result for the derivative to exist.
    But when we substitute for h->infinity, we only check the side h>0, right ?
    So shouldn't we also substitute u=-1/t and verify that we have the same result ?
    Or else prove that the derivative must exists in which case only one side is enough to get the value.

  • @mohammadpourheydarian5877
    @mohammadpourheydarian5877 2 года назад +1

    Smart moves and thank you. To avoid confusion in approach 1 instead of twice using u I will use U and then w.

  • @vegasuser999
    @vegasuser999 5 лет назад +5

    Wonderful videos. It is a long time ago that I studied complex variables, differential and integral calculus and algebra. So it is great fun watching this guy do with ease what most of us struggled with when learning the basic elements of these important mathematical techniques. I can generally follow him right to the end once I see where he headed. The mathematical manipulations seem to be firmly rooted in my brain. The algorithms he applies for problem solving are much less so.

  • @thomaskember4628
    @thomaskember4628 4 года назад +2

    This reminds me of when I was a COBOL programmer, we would have discussions about whether you could have positive zero and negative zero. This was because the sign of a number was contained in the units digit. So, when comparing numbers it was important to take this into account. But I would say to my colleagues that zero was neither positive nor negative, it was separate from other numbers.

  • @MH2029
    @MH2029 5 лет назад +4

    Hey , I recently started reading Thomas calculus and found that lnx was actually defined as definite integral of 1/t from 1 to x. So i think a proof is not needed stating the definition is enough. Anyway hats off to the great content

    • @carultch
      @carultch Год назад

      How it is defined, really depends on who you ask. Historically, natural log was discovered before the number e, and it was defined as this integral. But in modern times, we usually define it as the inverse of e^x, and define e^x as the special case of the exponential where it is its own derivative. The modern definition is much more useful, to learn what logs are for the first time.
      These two definitions are internally consistent, but you need to start with one to prove the other.

  • @BradleyG01
    @BradleyG01 2 года назад +1

    The most impressive thing about these videos is not the math, it's his ability to write with 2 or 3 markers in the same hand while holding them all at the same time. And that his writing is still legible while he does it. I can barely read my own handwriting when i write with just 1 pencil

    • @CliffSedge-nu5fv
      @CliffSedge-nu5fv Год назад

      And hold a microphone in the other hand. Might as well start juggling at that point.

  • @الأستاذطلعتالصفاوي

    جميل ورائع ومميز ما يقوم به هذا الشاب،،، فعلا عقليه فذه،، 🌹🌹🌹

  • @mathunt1130
    @mathunt1130 2 года назад +1

    You can do this in two ways. You can use the integral definition of log(c) and use the fundamental theorem of calculus or you can note that log(x) is the inverse function of exp(x), and just use the expression for differentiating the inverse function.

  • @elreturner1227
    @elreturner1227 5 месяцев назад

    The second lim going into the continues function was so eye opening and satisfying

  • @SaadmanSaif
    @SaadmanSaif 2 года назад +4

    Oh my god you are incredible! I learned a thing or two because of you! Loved it ❤️

  • @RYedukrishnan-cn5ft
    @RYedukrishnan-cn5ft 8 месяцев назад +1

    Hello sir ❤, thanks for this video.
    I have a doubt @4:03 min , why you put 1/h power their 😅. we know that ln(x)²≠ln(x²). You should put extra bracket there 😁I think. If I am wrong please inform me.
    My poor English hop you understand 😅.
    A fan student from India (leaving the exact place of Ramanujan)

  • @ClaudioCP
    @ClaudioCP 2 года назад

    Very elegant description of this important derivative

  • @itsflyin
    @itsflyin 4 года назад +4

    Thank YOU so much for sharing your beautiful smile and passion!! It makes me so much more excited to learn and genuinely happy :))

  • @ramblingwhitedog7346
    @ramblingwhitedog7346 11 месяцев назад

    Dang, when he finally pulled out the e term, I got super excited. Nice job!

  • @rodericksibelius8472
    @rodericksibelius8472 5 лет назад

    So clear explanation, Greatest Math Teacher in the WORLD, Thank You Sir!.

  • @xoppa09
    @xoppa09 Год назад

    I wonder if there is a numerical analysis class he teaches. This guy is a good teacher.

  • @elyseepasteur6163
    @elyseepasteur6163 5 лет назад +1

    Well done guy!
    You sort it out!
    Keep it up!
    Go always deep n in every detail to enlightening.
    Again you've done it!

  • @andreemery4964
    @andreemery4964 6 лет назад +1

    I think it could have been made a bit more clear at 3:29 that the 1/h exponent is supposed to be evaluated for (1+h/x) before the log is taken. (But I still got the point.)

  • @elreturner1227
    @elreturner1227 5 месяцев назад

    This was so intertwining I was guessing what to do and when he showed what to do it made sense feels amazing

  • @Syntaxxed
    @Syntaxxed 7 месяцев назад

    such a long proof but very well thought out. I was definitely doing a shorter proof for my test (luckily, not sure if I could survive writing this for my test.. lol).
    Dloga(x)=1/x*ln(a)
    D(log(e^x))=1/xlne=1/xloge(e)=1/x*1=1/x
    but of course mine is already making assumptions (that derivative of loga(x)=1/x*ln(a)) instead of figuring it out with definition of e. Great work, definitely I learned something.

  • @Velas99
    @Velas99 7 месяцев назад

    Just found your channel. Thanks for creating this content and keep up the good work.

  • @DaveJ6515
    @DaveJ6515 4 года назад +3

    Very well explained as usual, only one thing: I ask my students to avoid “canceling ln with e”; I want them to say that log of a power with the same base is the exponent.

    • @stewartzayat7526
      @stewartzayat7526 4 года назад +2

      I think that's just a linguistic thing. As long as your students know _why_ it works, I don't think it matters what they say

    • @DaveJ6515
      @DaveJ6515 4 года назад +4

      @@stewartzayat7526 Since that is the definition of a log, I prefer my students to repeat it as frequently as possible: it's the best way to capture it completely. It's part of my campaign against voodoo maths: you know, strange things like quantities that change their sign while flying over the equal sign and all that. An easy way to forget that there are equivalence principles behind that, and no flying stuff.
      Also: linguistics is a central part of our learning processes: our first impact with new stuff is via a language, so it makes lot of difference, imho.

    • @diegocabrales
      @diegocabrales 2 года назад

      Logarithm is the inverse function of exponentiation and viceversa, that's why log_a(a^x) = x and a^(log_a(x)) = x. I would prefer to say this rather than what you have written here (that includes the equivalent of what you have said for explaining that a^(log_a(x)) = x).

    • @DaveJ6515
      @DaveJ6515 2 года назад

      @@diegocabrales and of course you would be right, but my 36 years of experience teaching maths make me say that your students would benefit less from that explanation.

  • @frederickgriffth4431
    @frederickgriffth4431 3 месяца назад

    Excellent proof process! 中国人来RUclips看中国人用英语讲课了

  • @markharder3676
    @markharder3676 2 года назад

    I once tried to find a good thermodynamic definition of 'heat'. Every thermo text I have, and there are more than a handful, defines heat its own way. The most honest says that heat is what we measure with a calorimeter (Sort of like defining temperature as the thing we measure with a thermometer.). And what is a calorimeter? Why it's an instrument we use to measure heat of course.
    In the same spirit we can prove that the ln of x= the integral of 1/x. ( ln x = exp(integral(1/x dx) ). To do that, we must show that exp(integral(1/x dx) ) = x.
    Working out the first few terms of the Taylor expansion for the integral exponent will show that this equality is true. (I verified this with Mathematica. ) Since the definition of ln x is that it is the inverse function of the exponential: ln( exp(x) ) = x,
    we can substitute what we just proved for x in the definition: ln (exp (integral(1/x dx) )= integral(1/x dx). From the fundamental theorem of calculus (another inverse operation) we know that
    (ln (x) )' = d/dx (integral (1/x dx) = 1/x. QED
    I read Mat Hunt's answer and I think he and I are thinking the same thing. I just wanted to spell it out in full, except for that Taylor's series bit.

  • @perlindholm4129
    @perlindholm4129 2 года назад

    Idea - Solution to limit 1/x where x->0. Assume all derivatives have a defined initial value. dy_dx = 1/x. As x goes to zero there must exist a valid y(0) value. Just choose a limit such that it fits slope. Call it the perpendicular limit for 1/x. Or it is a freeze limit (black holes) when expressed in an equation. Depending on how fast you approach 0 by the 0.9 value you freeze the y values.
    import numpy as np
    # 1/x is a random number generator? when expressed in equation 0 = yx**2 - x
    x = np.random.rand(1000)
    y = np.random.rand(1000)
    while True:
    err = x**2*y - x
    x = 0.9*x + 0.001*err
    y = y + 0.001*err
    print(np.mean(err**2))

  • @mdghufranalam7369
    @mdghufranalam7369 6 лет назад

    Mst ...

  • @prabhamishra1099
    @prabhamishra1099 6 лет назад +1

    YOU ARE PHENOMENON!!!!!

  • @dhrubajyotisarma6268
    @dhrubajyotisarma6268 5 лет назад +1

    I like all the math problem and solutions 👍👍

  • @mokouf3
    @mokouf3 2 года назад

    I love u-sub when doing algebra and calculus. SO useful.

  • @JJCUBER
    @JJCUBER 5 лет назад +1

    Using the derivative of log, you can also see that since log_e(x) would be 1/(xln(e)), which simplifies to 1/x, we know that the derivative of lnx is 1/x.

    • @connorhorman
      @connorhorman 5 лет назад +1

      JJCUBER I do believe that the derivative of the generic logarithm comes from the derivative of the natural log, and the base change formula for logarithms, so you can’t use that.

    • @JJCUBER
      @JJCUBER 5 лет назад +1

      connor horman that wasn’t my point, I was saying that it works even for the derivative of lnx, for example if someone were to forget the derivative of lnx but not log_b(x)

    • @connorhorman
      @connorhorman 5 лет назад

      JJCUBER I suppose its useful if you are just trying to remember derivatives. It can’t be used in a proof, however.

    • @JJCUBER
      @JJCUBER 5 лет назад +1

      connor horman well using it as a proof wasn’t my goal

  • @Epic-fb8pe
    @Epic-fb8pe 3 года назад

    thanks brother for clear information ❤🤗🤗😎😎

  • @plantita3183
    @plantita3183 2 года назад

    THANK YOU!!! All other videos I found only explained how you used the derivative not actually showing proof on why it’s 1/x

  • @رضاشریعت
    @رضاشریعت 5 лет назад +2

    Love the second proof of lnx's derivative

  • @ddg-norysq1464
    @ddg-norysq1464 4 года назад +2

    Wait, doesn't the power 1/h belong inside the brackets?

  • @dalenassar9152
    @dalenassar9152 5 лет назад

    I just wanted to say, that for some reason, LOGb(X)=ln(X)/ln(b) has always been my favorite relationship in "Logarithmic Functions" and THANKS for the bonus at the end!!!

    • @jamesfortune243
      @jamesfortune243 2 года назад

      X = b ^ logb(x), then take logd of both sides and bring the exponent down. Then solve for logb(x).

  • @blazep5881
    @blazep5881 7 лет назад +35

    Wasn't this video already posted once?? I remember seeing it

    • @eliascaeiro5439
      @eliascaeiro5439 7 лет назад +4

      Yes but it got deleted for some reason.

    • @blackpenredpen
      @blackpenredpen  7 лет назад +13

      Pranav B yes. But I forgot to consider cases in that previous vid. And also this is a better pf

    • @koenzeven
      @koenzeven 7 лет назад

      I think it's a mix of 2 videos

    • @blackpenredpen
      @blackpenredpen  7 лет назад +6

      Oh btw, I did add a bonus part at the end lol

    • @RoyEduworks
      @RoyEduworks 6 лет назад

      ruclips.net/video/XYYdl9A__mo/видео.html

  • @raphaelmillion
    @raphaelmillion 7 лет назад +1

    Very easy method:
    f(x)=x=e^ln(x)
    f'(x)=1=ln'(x)*e^ln(x)
    1=ln'(x)*x
    ln'(x)=1/x
    Qed

  • @NeonArtzMotionDesigns
    @NeonArtzMotionDesigns 6 лет назад +2

    Before watching:
    Oh! What a coincidence, my calculus class just did that
    After watching: Oh! What a coincidence, my calculus class just did

  • @Chris89892
    @Chris89892 10 месяцев назад

    A most elegant solution to d/dx Ln(x)...I didn't imagine it would take 3 substitutions.

  • @libelldrian173
    @libelldrian173 3 года назад +2

    The closer he gets to 1/x, the bigger his smile becomes. :D

  • @marsbrook1
    @marsbrook1 3 года назад

    Absolutely beautiful. Great explanations! Thank you.

  • @atikhasan3027
    @atikhasan3027 3 года назад

    Awesome explanation

  • @AnthonyPerez-ls7sq
    @AnthonyPerez-ls7sq 6 лет назад

    I love you, plain and simple.

  • @bhavydugar6665
    @bhavydugar6665 3 года назад +1

    At the third step you could have broken the fraction into 1+h/x and then divided and multiplied the base h with x instead of using it is as power then you would have got 1/x lim h->0 ln(1+h/x) divided by h/x and then by using the limit you could’ve simply got 1/x

  • @ZehcnasEsined
    @ZehcnasEsined 6 лет назад

    It's so beautiful ♥
    Great explanation!

  • @douglasmagowan2709
    @douglasmagowan2709 2 года назад +2

    Back when I learned this we defined the logarithm function in terms of the integral 1/x dx, then proved that this function had the properties expected of a logarithm.

    • @luisvasquez-ib1dk
      @luisvasquez-ib1dk Год назад

      CIERTO BRO YO TAMBIEN LO APRENDI AL REVES,QUE EL LOGARITMO SE DEFINE JUSTO POR LA INTEGRAL DE 1/X

  • @kruksog
    @kruksog 5 лет назад +7

    Was it proved when this was discovered that the limit of a continuous function is said function of the limit? I know in the beginning, calculus was not super rigorous. And if not... how did they figure out the derivative of ln(x)?

    • @charbeleid193
      @charbeleid193 5 лет назад +2

      The prood might not be but that's actually quite an intuitive statement. And if I find it intuitive, someone like napier or leibniz definitely also did...

    • @me_hanics
      @me_hanics 4 года назад

      they did it by definition, and the definition already existed before

    • @98danielray
      @98danielray 4 года назад

      @@me_hanics you need the fact the function is continuous for that

  • @gnikola2013
    @gnikola2013 7 лет назад +5

    But bprp, there is a problem! lim x-> 0 [(1+u)^(1/u)]^(1/x) isn't exactly a continious function. The limit is equal to e, then the function would be e^(1/x), whose domain doesn't include 0. So technically it is just continious only for the values of x that you are interested in, as in the original function 1/x, the domain also excludes 0, so there is no good in trying to find out the derivative in that point. Just a technical detail but it's important isn't it?

    • @achyuthramachandran2189
      @achyuthramachandran2189 6 лет назад +1

      Kiritsu it's limit u->0, not x... Remember, when u do the substitution, u change the limit too

    • @angelmendez-rivera351
      @angelmendez-rivera351 5 лет назад

      The domain does not need to include 0 for it to be continuous in the neighborhood of x = 0. Remember, we are evaluating the limit of the value of the function as x APPROACHES 0, NOT the value of the function AT x = 0. The function is continuous for all nonzero values of x. Therefore, the limit and the function commute and "can be interchanged".

  • @michelkhoury1470
    @michelkhoury1470 2 года назад

    Another method:
    Let f(x) = exp (x)
    We have f'(x) = f(x) then f'(x)/f(x) = 1
    And ln(f(x)) = x
    So (ln(f(x))' = (x)' = 1 = f'(x)/f(x)
    Finally, let f(x) = x to find (ln(x))'

  • @igt3928
    @igt3928 6 лет назад +8

    1 to the infinity power is 1, 1 plus an infinitesimal value, all to the infinity power, could be anything.

    • @dp-zn8bd
      @dp-zn8bd 6 лет назад +1

      Except it converges to e

    • @igt3928
      @igt3928 6 лет назад +2

      lim n->inf (1)^n=1
      lim n->inf (1+Ln(k)/n)^n = k
      That's what I was saying, sorry if it wasn't clear. It's just that he says "one to the infinity power" and that's one, every time, but "one plus an infinitesimal bit", that's another story

    • @Paul-ob2hy
      @Paul-ob2hy 5 лет назад

      Ignacio but since when does 1/0=infinity? am i being stupid or something but i swear it isn’t defined

    • @angelmendez-rivera351
      @angelmendez-rivera351 5 лет назад +1

      Ignacio Except you are wrong. Saying that he was supposed to say "one plus one infinitesimal bit" is ridiculously ignorant. There is no such a thing as "an infinitesimal bit". There are no infinitesimal numbers. The real numbers contain no infinitesimals. 1^♾ is universally well-known and well-defined as an indeterminate form of limits. This is what BPRP was referring to.

    • @justabunga1
      @justabunga1 5 лет назад

      Just because 1 raised to any power is 1, infinity is not a number. 1 raised to infinity power is indeterminate. We have to use l'Hopital's rule to do more work.

  • @JayTemple
    @JayTemple 2 года назад +3

    Which was proven first, the derivative of e^x or that of ln x?

  • @GeofreySanders
    @GeofreySanders 2 года назад +1

    That moment when a channel about two colors of pen PULLS OUT A THIRD COLOR

  • @Zonnymaka
    @Zonnymaka 7 лет назад +32

    Here's a mind blowing question:
    Is Log_2_(3) bigger/smaller/equal than/to Log_3_(5)?
    That's a so called "coffin" problem, in other words that's one of the questions that russian teachers were use to ask to a russian jew in order to prevent him/her to enter the math faculty (no kidding!)

    • @blackpenredpen
      @blackpenredpen  7 лет назад +10

      Zonnymaka wow!!! Super interesting problem. I will have to think hard on it

    • @Zonnymaka
      @Zonnymaka 7 лет назад +3

      How is it going? :)
      Tip: don't waste you time (as i did) comparing the 2 logs...you'll never get out alive of that!

    • @blackpenredpen
      @blackpenredpen  7 лет назад +3

      Zonnymaka my initial try was to set up a function but no luck. I will have more time tmr or so to think about it.

    • @raphaelmillion
      @raphaelmillion 7 лет назад +2

      got this pretty quicky, hope this is correct:
      log_2(3)=x and log_3(5)=y
      thus 2^x=3 and 3^y=5
      setting these equal we get:
      2^x=3/5*3^y
      now take the log2 again:
      x=log_2(3/5)+y*log_2(3)
      we have a function f(y) = x
      set up new function f*(y)=f(y)-y
      note that if this is positive, the x value is greater than the y value and vice versa
      f*(y)=log_2(3/5)+y*(log_2(3)-1)
      =log_2(3/5)+y*log_2(3/2)
      the first part is positive since the inside is greater than 1
      the second part is positive because y is positive and the inside of the log is also greater than 1.
      Since 0

    • @raphaelmillion
      @raphaelmillion 7 лет назад +1

      just looked it up on wolframAlpha, seems to be correct

  • @TheLifeLaVita
    @TheLifeLaVita Год назад

    1/ln(b) can also be written as logb(e) so you don't have it under a fraction 👍
    (the derivative of a general logb(x) function would be (1/x) * logb(e))

  • @merveilmeok2416
    @merveilmeok2416 5 лет назад

    Wonderful work!

  • @sorprenant9106
    @sorprenant9106 2 года назад +1

    I never understood why people said math was beautiful until I got to college and started calculus.

  • @donmoore7785
    @donmoore7785 4 года назад

    Excellent, as usual!

  • @chilliblogs9350
    @chilliblogs9350 7 лет назад

    I watch your videos for inspiration and help as I just started year 7

  • @charbeleid193
    @charbeleid193 5 лет назад +1

    You know what's the best thing about maths? It's that they work

  • @hpholland
    @hpholland 2 года назад +3

    Let’s see how good your math skills are when you put down your ball of power

  • @tarikabdelhadibenaouda
    @tarikabdelhadibenaouda 2 года назад

    the first definition are amazing

  • @yamomzameanskeet
    @yamomzameanskeet 5 лет назад

    The limit definition of the derivative of ln(x) is a nice one!

  • @ukidding
    @ukidding 4 года назад

    well done professor

  • @basil9633
    @basil9633 Год назад +1

    I appreciate the derivative 1/x now

  • @lucasmansson1055
    @lucasmansson1055 Год назад

    Proving this was actually a question on one of our calc exams

  • @alexiapineda6519
    @alexiapineda6519 6 лет назад

    Thank you, you are the best explaining ♥️

  • @PuzzleQodec
    @PuzzleQodec 7 лет назад +1

    Here's what I tried, working backwards from a proof of the chain rule, and using this alternative definition of e (see also ruclips.net/video/SxJ7X8vE-f0/видео.html):
    1 = lim[u->0] (e^u - 1) / u
    From which I concluded (but I wonder if this is correct):
    1 = lim[u->0] u / (e^u - 1)
    Given x > 0, by definition:
    ln′(x) = lim[h->0] 1/h ∙ (ln(x + h) - ln(x))
    = lim[h->0] 1/h ∙ ln((x + h) / x)
    Now let u = ln((x + h) / x), which is defined for x > 0 and h sufficiently small.
    -> if h->0, then u->ln((x + 0) / x) = ln(1) = 0
    -> u = ln((x + h) / x)
    => e^u = (x + h) / x
    => x∙e^u = x + h
    => x∙e^u - x = h
    Going back and substituting:
    ln′(x) = lim[h->0] 1/h ∙ ln((x + h) / x)
    = lim[u->0] 1/h ∙ u
    = lim[u->0] 1/(x∙e^u - x) ∙ u
    = lim[u->0] 1/x ∙ 1/(e^u - 1) ∙ u
    = 1/x ∙ lim[u->0] 1/(e^u - 1) ∙ u
    = 1/x ∙ lim[u->0] u / (e^u - 1)
    = 1/x
    Feels like this alternative definition of e is tailored exactly for deriving derivatives. ;-)
    *Edit*: added link to your recent video about e.

  • @shivanshnigam4015
    @shivanshnigam4015 Год назад

    Lim x->0[ln(1+x)/x] = 1
    A standard limit can also be used