See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/348007076

Negative Factorial

Article  in  International Journal of Mathematics and Computer Science · September 2018

DOI: 10.47191/ijmcr/v8i12.01

CITATIONS READS
0 698

1 author:

Phani Srikar
Sreenidhi Institute of Science & Technology
1 PUBLICATION   0 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Negative Factorial View project

All content following this page was uploaded by Phani Srikar on 31 December 2020.

The user has requested enhancement of the downloaded file.

 Negative Factorial
Phani Srikar1
1
Department of Electronics and Communication Engineering, Sreenidhi Institute of Science
and Technology,JNTUH University
[email protected]

10 August 2020

Abstract 2 History

The main focus of this paper is to present a way to

think bout the existence of negative factorials and
how to extend the idea of non-positive factorials ,and
stress more on the thought process of realising their
existence rather than their application ,and under-
stand deeply about the origin of the idea and then Multiple scientists have worked on this subject ,but
analyse their functional behaviour ,using the LIMIT- the principal inventors are those who gives the
ING approach. And focus on the application using asymptotic formula after some work in collaboration
the idea of damped oscillations towards increasing the with De Moivre, finally and introduces the actual no-
stability of high altitude bridges and apply it to its tation n!.Of course other scientists such as Taylor
electrical analogy. have also worked a lot with this notation.The no-
tation n! was introduced by French mathematician
Christian Kramp(1760-1826) in 1808.The term fac-
torial was first coined in French as factorielle by an-
other French mathematician Louis Francois Antoine
Arbogast(1759-1803) and Kramp decided to use the
1 INTRODUCTION term factorial so as to circumvent printing difficul-
ties incurred by the previous used symbol Ln. (?,
?)gave the additive representation of factorials. (?,
Factorials are a magical discovery in the field of math- ?) gave an expository account of the factorials, gave
ematics.Their use has always been in the fields of several new results and posed certain problems on
combinatorics ,calculus ,probability right from their factorials. Research on the interpolation of factorials
discovery and the actual behaviour of them has been started with correspondence among Leonhard Euler,
neglected.We have always stressed on their animal Daniel Bernoulli and Christian Goldbach in the year
application rather than their magical behaviour and 1729 (Refer to the correspondence reproduced by (?,
left us blinded by their applicative properties and we ?)). (?, ?) gave an account of the early history of the
have forgotten to experiment with their functional factorial function. (?, ?, ?) defined the factorial of
behaviour. negative integers until a certain extent.

1
3 The Origin Of Thought 3.2 Analogy to Factorial
Now this proof has made me wonder and inspired to
3.1 Realising the idea of LIMITING approach the proof of 0! in a similar manner.
approach with the aid of a0 proof
n! = n(n − 1)(n − 2).....1.
In mathematics ,the factorial of a non-negative inte-
ger n, denoted by n!, is the product of all positive
integers less than or equal to n.
For example : By definition of factorial ;

n! = n(n − 1)(n − 2).....1 (1)

The value of 0! is 1, according to the convention

for a an empty product.

It all began with the proof of a0 = 1,where a is in

any real number = 0.

Figure 2: LIMITING approach to prove 0!

Here we have divided the subsequent factorial value

by the factorial number in a descending way and ap-
proach towards 0 and hence 0! is found equal to 1.
Now just like a0 proof what if we try to simulate the
existence of factorial of non-negative numbers using
the LIMITING approach in such a way that they seem
to approach 0!.

3.3 Failure of Gamma Function exten-

sion to negative numbers
By referring to gamma function (?, ?) and (?, ?)
we have Γ(n) = (n − 1)! ,for n > 0,where n ∈ Z.
Figure 1: LIMITING approach through the proof of
Legendre in 1808 gave the notation to the Euler’s
a0
gamma function (?, ?).

It is clear by definition that gamma function fails

This is somewhat similar to a LIMITING ap- for non-positive numbers and it value tends to infinity
proach towards a0 = 1, where we try to approach (?, ?, ?).
from both the sides towards a0 by symmetric math-
ematical operations. i.e. here we multiply by a for
3.4 Redefining Gamma Function
negative powers of a and divide by a for positive pow-
ers ,where powers from positive and negative numbers But what if we increase the values from a given
tend to reach 0. negative number and make it approach 0!, this is just

2
opposite to what we do with factorials of positive One can also compare the functional values of n!
numbers , just like we decrease positive numbers by and (−n)! and see that the values are same in mag-
1 we increase the negative numbers by 1. nitude but differ in their sign alternatively, i.e. neg-
ative factorial of odd numbers bear the opposite sing
This could potentially lead to some new defini- as that of the factorial of positive odd numbers ,but
tion for extension of factorials to negative numbers. the negative factorial of even numbers are same as
that of factorials of even numbers. Among the other
This is also just like the limiting approach where well defined functions for the factorials of real nega-
we tend to reach 0! From positive side by decreasing tive numbers are, Hadamard’s gamma function (?, ?)
values ,and for the negative side in order to show and Luschny’s factorial function (?, ?), both of which
such similar behaviour , the only thing we could do are continuous and positive at all real numbers.
to them make approach 0! is to increase them i.e add
1 rather subtracting 1 from them just as the gamma Now this how the LIMITING proof of 0! looks as
function does which fails to halt. shown in Figure 3.

For example , Lets us try to attempt to increase

the negative numbers from a given number to ap-
proach 0! ,

(−5)! = −5 · −4 · −3 · −2 · −1 = −120
Here we stop at -1 ,as extending further is of no use
as our primary objective was to just approach until
0! .

(−4)! = −4 · −3 · −2 · −1 = +24
(−3)! = −3 · −2 · −1 = −6 Figure 3: Limiting approach to prove 0! = 1
(−2)! = −2 · −1 = 2
(−1)! = −1
3.5 Conclusion
By practical observation and rom the approach
The results as shown in Figure 3 support our for-
that we have decided we can formulate this behaviour
mula by showing a symmetrical approach (just like
as
a0 proof) towards 0!.

(−n)! = (0 − n) · (1 − n) · (2 − n) · ......... − 1.

or simplifying this we get

(−n)! = (−1)n · n!.

4 Graphical Analysis
Here this formula completely justifies our vision of
increasing the negative numbers to approach 0!.Just while searching for the existence or validation
like the positive numbers decrease themselves to ap- of the negative factorial by some real world func-
proach 0!. tion(existing in present without any ambiguity) ,the

3
discrete valued graph of the function f (t) = e−at · 4.1 Findiing the values of the con-
sin(bt) (?, ?)seems to have some similarity with the stants a, b
graph of (−n)!.
4.1.1 Trial and Error method
The results of the following graphs plotted using But for what values of the constants in the func-
MATLAB show some encouraging results. tion (a, b, A) would this be possible and what should
be nature of the amplitude, constant or variable? well
first let’s see for what values of a and b the function
varies in the similar manner to negative factorial
In MATLAB I plotted the graph of the function
for negative values of t(for the function) and n (for
its discrete form).

the graphs for a, b > 0 are as follows:

Figure 4: Graph of f (t) = e−at · sin(bt) in discrete

and continuous domain

Now let us compare this with the discrete graph of

−n! according to out our definition that we formu-
lated.

Figure 6: graphs of f (t) for a, b > 0

well these graphs does not seem similar to nega-

tive factorial graphs. Hence next I choose the values
of a as0 < a < 1 and b > 0, now the results seem
promising
Figure 5: negative factorial graph for negative ac-
ceptable values in its domain I kept changing the values of a from 0.1 to 0.4 and
b value from 1 to 4 but the right combination was
One can clearly observe that the variation of the a=0.3 and b = 3 that provided the right frequency
the discrete graph of the function and negative fac- as that of the negative factorial graph.
torial is similar and only vary in amplitude and can close experimentation and multiple trials has found
be easily corrected to match the value of negative that the constants depend on pi i.e. a = (pi - (1/pi2)
factorial at that point. /10) = 0.30403 and b = pi - (1/pi2) = 3.0403.

4
 Figure 8: N vs Amplitude
Figure 7: graphs of f (t) for 0 < a < 1 and b > 0

4.1.2 Interpolating the amplitude 3 \

4 a = ( pi − (1/( pi ∗ pi ) ) ) /10;\
Now coming to amplitude ,it can be chosen in such 5 b = pi − (1/( pi ∗ pi ) ) ;\
a way that the function value attains the value of 6 \
negative factorial at that particular value of -n . 7 %h e r e we p l o t t h e graph f o r A5 ( f i x e d
Since we know the value of (-n)! we can equate it amplitude ) \
to the function and find for what value of amplitude 8 z = −5;\
the function can be equal to the value of negative 9 z1 = exp(−a ∗ z ) ; \
factorial(?, ?) at that point. 10 z2 = s i n ( b∗ z ) ; \
since we have An*e-an *sin(bn) = (-n)! 11 z3 = abs ( z1 . ∗ z2 ) ; \
Hence An = mod((-n)!/e-an*sin(bn)) ,where mod12 A = f a c t o r i a l (−z ) / z3 ; \
indicates the modulus function i.e mod(x) = x and13 \
mod(-x) = x. 14 n = −5:0;\
Here is a table showing the value of An vs n to15 n1 = exp(−a ∗n ) ; \
help the function reach the value of (-n)! All of16 n2 = A. ∗ s i n ( b∗n ) ; \
these A values were found using he expression An17 n3 = n1 . ∗ n2 ; \
= mod(n!/e-an*sin(bn)) and n = (-n). Where mod18 stem ( n , b1 ) ; \
is the modulus function. 19 h o l d on\
20 stem ( n , n3 ) ; \
Now let us choose a value of A and plot the graph21 h o l d on\
to compare with negative factorial graph. Using the22 \
following code in MATLAB we can plot the graph23 t = −5:0.01:0;\
upto n = -5 and choosing the appropriate A. 24 y = A. ∗ s i n ( b∗ t ) ; \
25 y2 = exp(−a ∗ t ) ; \
26 y3 = y2 . ∗ y ; \
4.2 MATLAB code for comparing the27 p l o t ( t , y3 ) ; \
function and −n! 28 \
29 \
1 b2 = 5 : − 1 : 0 ; \ 30 h o l d on\
2 b1 = (( −1) . ˆ b2 ) . ∗ f a c t o r i a l ( b2 ) ; \ 31 p l o t ( n , b1 , ’ r ’ , n , n3 , ’ g ’ , t , y3 , ’ b ’ ) ; \

5
32 } up-to a certain -n. It represents all the 3 graphs
(f(t) its discrete(green) , continuous form(blue,and
the negative factorial discrete plot(red impulses))

4.3.1 MATLAB code for Interpolating and

plotting the graphs

2 clear all ;
3 close all ;
4 clc ;
5 number = i n p u t ( ’ e n t e r t h e number : ’ ) ;
6 b2 = −number : − 1 : 0 ;
7 b1 = (( −1) . ˆ b2 ) . ∗ f a c t o r i a l ( b2 ) ;
8

Figure 9: NFactComparision with function 9 a = ( pi − (1/( pi ∗ pi ) ) ) /10;

10 b = pi − (1/( pi ∗ pi ) ) ;
11 z = −1: −1:number ;
The blue curve represents the function f(n) in a12 z1 = exp(−a ∗ z ) ;
continuous domain and the green impulses are the13 z2 = s i n ( b∗ z ) ;
discrete mappings of the negative factorial and red14 z3 = abs ( z1 . ∗ z2 ) ;
impulses represent the discrete form of the continuous15 f o r i = 1:−number
function f(t). 16 z4 ( i ) = f a c t o r i a l ( i ) / z3 ( i ) ;
We can observe that at n = -5 both function value17
and negative factorial value are equal this happened18 end
because of the value of the amplitude A that we have19 x=1:1: −number ;
chosen but we can also see that the function does not20 y=z4 ;
equate to the values of negative factorial at points n21 sum=0;
= -4,-3,-2 and -1. 22 f o r i =1: l e n g t h ( x )
Hence to make all the values of negative factorial23 p=1;
lie on the curve f(t) we can use the amplitudes that24 f o r j =1: l e n g t h ( x )
would make this possible at their respective -n and25 i f j ˜= i
use Lagrange’s interpolation to find A as a function of26 c = p o l y ( x ( j ) ) / ( x ( i )−x ( j ) )
n ,say A(n) where it satisfies A(-1) = mod(A-1) ,A(- ;
2) = mod(A-2),A(-3) = mod(A-3), . . . . and A(-n) =27 p = conv ( p , c ) ;
mod(A-n) . 28 end
29 end
4.3 Interpolating and plotting the30 term = p∗y ( i ) ;
31 sum= sum + term ;
graphs
32 end
I plotted the graph by finding A as a polynomial of n33 d i s p ( sum ) ;
and satisfying A(i) for i = - 1,-2,-3,-4,-5,-6 in MAT-34 p o l y = poly2sym ( sym ( sum ) ) ;
LAB and we can see that all the values of -n! for n¡=35 disp ( poly ) ;
-6 lie on the curve . 36 n = −number : − 1 : 0 ;
The code for this is below which takes input from37 answ = ( s u b s ( poly , n ) ) ;
user to plot the graphs from the above rules to plot38 A = answ ;

6
39 N = number : 0 ;
40 N1 = exp(−a ∗N) ;
41 N2 = A. ∗ s i n ( b∗N) ;
42 N3 = N1 . ∗ N2 ;
43 stem (N, b1 ) ;
44 h o l d on
45 stem (N, N3) ;
46 h o l d on
47 y = −number : − 0 . 0 1 : 0 ;
48 A2 = s u b s ( poly , y ) ;
49 Figure 10: Interpolated negative Factorial
50 t = number : 0 . 0 1 : 0 ;
51 y = A2 . ∗ s i n ( b∗ t ) ; Now we can clearly see that all the values of nega-
52 y2 = exp(−a ∗ t ) ; tive factorial perfectly lie on the curve ,we can clearly
53 y3 = y2 . ∗ y ; say that we have found a perfect function equivalent
54 p l o t ( t , y3 ) ; for the negative factorial that we defined according
55 h o l d on for their existence.
56

57 p l o t (N, b1 , ’ r ’ ,N, N3 , ’ g ’ , t , y3 , ’ b ’ ) ;
5 Conclusion
As we have seen from the above discussions and proof
that negative factorials exist and show very mystical
behaviour both in their mathematical and graphical
analysis.
Negative factorial of a number n can be defined
as the product of numbers greater than -n up to -
1 . We have seen that factorials of non-negative and
non-positive numbers have same values with different
signs.
As we have also seen the graph on -n vs (-n)! Is
found to be damped oscillatory motion.
Hence the graph of negative factorial is similar to
that of the function f (t) = e − at ∗ A ∗ sin(bt) for
particular values of a, b, A i.e. a = (pi - (1/pi2) /10)
= 0.30403 and b = pi - (1/pi2) = 3.0403. and A is
a Lagrange’s interpolated polynomial A(n) of degree
n-1 found using -1,-2,-3. . . ..-n as x values and A(-1)
= mod(A-1) ,A(-2) = mod(A-2),A(-3) = mod(A-3),
. . . . and A(-n) = mod(A-n) as the corresponding y
values where An = mod((-n)!/e-an*sin(bn)) ,where
mod indicates the modulus function i.e mod(x) = x
and mod(-x) = x and n = -1,-2,-3,,,,,,-n.
Hence we can clearly state that negative factorials
exist (according to the definition that we have given)
and are found to be equal to a function f(t) = e-at *

7
 A*sin(bt) for particular values of a, b, A. Oldham, K. B., Myland, J., & Spanier, J. (2010).
We have also understood how to think or develop a An atlas of functions: with equator, the atlas
thought process to arrive at the existence of factorials function calculator. Springer Science & Busi-
for negative numbers at the same time understating ness Media.
0 Thukral, A. K. (2014). Factorials of real nega-
the proof of 0! And a = 1 ,a is any real number not
equal 0 though the idea of limiting approach towards tive and imaginary numbers-a new perspective.
a number from either sides. SpringerPlus, 3 (1), 658.

References
Anglani, R., & Barile, M. (2007). Factorials as sums.
arXiv preprint math/0702010 .
Bhargava, M. (2000). The factorial function and
generalizations. The American Mathematical
Monthly, 107 (9), 783–799. Retrieved from
http://www.jstor.org/stable/2695734
Borwein, J. M., & Corless, R. M. (2018). Gamma
and factorial in the monthly. The American
Mathematical Monthly, 125 (5), 400–424.
Davis, P. J. (1959). Leonhard euler’s integral: A his-
torical profile of the gamma function: In memo-
riam: Milton abramowitz. The American Math-
ematical Monthly, 66 (10), 849–869.
Dutka, J. (1991). The early history of the factorial
function. Archive for history of exact sciences,
43 (3), 225–249.
Gautschi, W. (2008). Leonhard euler: his life, the
man, and his works. SIAM review , 50 (1), 3–
33.
Ibrahim, A. (2013). Extension of factorial concept
to negative numbers. Notes on Number Theory
and Discrete Mathematics, 19 (2), 30–42.
Lefort, X. (2002). History of the logarithms: an ex-
ample of the development of a concept in math-
ematics. Project Penelope, editor: Miguel Her-
nandez Gonzalez) Fundaciõn Canaria Orotava
de Historia de la Ciencia, La Orotava, 142–
155.
Lexa, M. (2002). Remembering john napier and his
logarithms. IEEE Potentials Magazine.
Loeb, D. E. (1995). A generalization of the binomial
coefficients. arXiv preprint math/9502218 .
Luschny, P. (2012). Approximation formulas for the
factorial function n.

View publication stats