Beruflich Dokumente
Kultur Dokumente
John Gough — jugh@deakin.edu.au — revised 21 November 2005
An online mathematics teacher, Rex Boggs (rboggs@bigpond.net.au) asked the AAMT
(Association of Mathematics Teachers) online interest group (Boggs: 20 Mar 2003):
Does anyone teach (or have an opinion on) the common denominator method of
dividing fractions, possibly introduced as follows (using 'divide' for the divide sign)?
For example, if we ask:
— What fraction of 4 apples is 3 apples?
… the easy answer is — 3/4 of one apple.
Then if we ask an equivalent fractionbased question:
— What fraction of 4 fifths is 3 fifths?
… we will have the same answer — 3/4 — three parts out of the four parts we have at
the start.
That is, 3/5 divided by 4/5 = 3/4.
The mathematical setting out of a more complex problem would be:
3/4 divided by 5/6
= 9/12 divided 10/12
= 9/10
In this case, dividing quarters by sixths, we resorted to using a common denominator
(twelfths) to establish a direct comparison between the two fractions.
What are its strengths and weaknesses, compared to the old rote advice:
“Yours is not to reason why, just invert and multiply” (which is a catchy parodyversion
of a line from Tennyson’s poem “Charge of the Light Brigade”).
that so many of us cherish.
If you do use it, how do you do so?
Rex Boggs
I replied with a version of the following discussion. I have revised and improved this
(November 2005).
Dear Rex, and others,
I'll confess I am not sure I follow the explanation you've given here.
For one thing, I can't see how we get from Rex's opening question "What is the common
denominator method of dividing fractions?" to the next question "What fraction of 4
apples is 3 apples?" How does (3/5) ÷ (4/5) relate to "What fraction of 4 apples is 3
apples?"
Maybe what Rex has in mind is that we compare the division of whole numbers (3 ÷ 4)
with a similar division of FIFTHparts of the whole numbers (3fifths divided by four
fifths), and use the relative sizes to argue that the result is the same.
Consider, amongst other possibilities D.A. Johnson and G.R. Rising Guidelines for
Teaching Mathematics, Wadsworth, Belmont, 1967; 2nd edition 1972 pp 146147 —
"Division by a rational number".
The discussion by Johnson and Rising mentions that "changing fractions [when they do
not having common denominators] to fractions with common denominators" is one easy
strategy for justifying the rote rule of invertandmultiply.
My next comments may therefore be irrelevant, but I won't let that stop me, in case they
are useful.
If we are talking about one fraction dividing another, I like to go back to the actual
process of division as it works with whole numbers, e.g. 14 ÷ 3.
Dividing (by a whole number) is either:
• sharing, or dealing out — e.g. share 14 objects between 3 people;
or
• grouping, or finding out how many nlots of things we can make using mnumber of
things — e.g. how many lots of 3 objects we can make with 14 objects.
Then I comment that any attempt to share 3/5 of an object between 4/5 of a person is
absurd (as you need at least one person to "share"), so we don't even try to think about
dividing fractions in a "sharing" way.
(However, if we interpose the almostabsurd idea that "When ONE person can share 3/5
of an object, and get all of the 3/5", we can then ask how much of the 3/5 would 4/5 of
that ONE person get, in which the case the answer would be 4/5ths of 3/5 and the
missing 1/5 of the ONE person would get the remaining unshared part of the initial 3/5
— OK?)
So instead of trying to work with divisionassharing when we consider fractions, I
propose working with divisionashowmanylotsof?
This means that I read a symbolstatement such as
— 3/5 ÷ 4/5
as
— How many lots of 4/5 are there in 3/5?
(Unfortunately, the similarity of denominators may confuse us. And it occurs, again, to
wonder how such an example fits with Rex's question about a "common denominator
method", when the two fractions in question start out already having the same
denominator. I tend to think of a "common denominator method" as being whatever you
have to do to concoct or find a helpful common denominator when the two fractions
being considered DON’T start out with the same denominator.)
This becomes a little easier to think about if we make a diagram that shows a WHOLE,
and also shows 3/5, and then another diagram that shows a WHOLE and also shows 4/5.
This is the WHOLE, shown with FIVE equal parts.
This is 3/5 of the WHOLE
and
shows 4/5 of the whole.
We can then ask, and visually consider, how many lots of 4/5 are there in 3/5;
or, equivalently,
— how MUCH of the 4/5 diagram is there in the 3/5 diagram?
Expressed in this last way it is easier to see that the 3/5 diagram has THREE out of the
FOUR parts of the 4/5 diagram.
This visual situation can also be expressed by saying that:
— 3/5 is actually threequarters of 4/5.
Try it: look at the diagram of 4/5, and work out what threequarters of this diagram is.
The answer is clearly 3/5.
Putting this another way, we can literally SEE that 3/4 OF 4/5 is 3/5 —
or
A threequarter portion of 4/5 = 3/5.
Remember that “taking portions”, like “so many lots OF some quantity”, is equivalent to
the process of multiplying (putting equalsized groups of things together, and finding
the total number of things in the combined groups).
Hence 3/4 of 4/5 = 3/4 x 4/5 = what is a threequarter part OF fourfifths?
Or, expressing this slightly differently and in a rather odd wording (the standard
language of “multiplication” as a process), the answer to the question "how many lots of
4/5 are there in 3/5?" is that there is 3/4, or, there is 3/4 of the initial 4/5.
By providing enough experience and lost of practical examples of this, a student may
realise the general pattern implicit in the diagrams: that is, we are eliminating
(cancelling) the common denominator.
At the same time, we are putting the uppernumber (numerator) of the divisor (the
fraction doing the dividing), into the lowernumber (denominator) of the dividend (the
answer at the end of the dividing).
In short, we infer (that is, we decide that we have evidence for a general rule or pattern
that describes ALL the examples we have so far investigated) that we are, as the rote rule
tells, us to do, inverting and multiplying.
Now consider a different case.
When the two fractions we begin with do NOT both have a common denominator it is
impossible (or not immediately possible) to find or see the relationship between the
fraction doing the dividing, and the fraction being divided.
Again I find that some diagrams help explain the idea, and also model the actual making
of a common denominator (BUT note that this may not necessarily be the LOWEST
common denominator).
Consider, for example, 2/3 ÷ 4/5.
How many lots of 4/5 are there in 2/3?
Or
How much of 4/5 is there in 2/4?
This time, instead of using linechunks, or “fractionbars”, as models for WHOLE and
PART(S) OF WHOLE I find it helpful to use a SQUARE as the model for a WHOLE.
Then we use either COLUMNS (or ROWS) as models for the PARTS of the WHOLE.
In particular I start by showing the number being divided as vertical columns within the
WHOLE SQUARE. Then, for instructional contrast, I show the number and do the
dividing in terms of horizontal rows, within the WHOLE.
So a WHOLE (or a unit, or ONE, or 1) looks like this
Obviously 2/3 looks like this — the three lefthand columns (the TWO thirds) have been
shaded (with slanted lines, from bottomleft upwards to topright) for visual emphasis,
leaving the UNUSED part of the whole UNSHADED — don’t let this confuse you —
the FRACTION is the shaded part of the diagram:
and the 4/5 looks like this (note that the shading slants in the opposite way, topleft
down to bottomright):
When we ask the question "How many 4/5 are there in 2/3?" it is not immediately
visually obvious.
But the next trick with diagrams helps us.
The visual difficulty is that we can't see how to compare the ///shaded THIRDs
columns with the \\\ shaded FIFTHs rows. But this becomes clear if we crossrule the
vertical columns in the 2/3 diagram so that the diagram now shows the horizontal lines
for FIFTHS, and we cross rule the horizontal rows in the 4/5 diagram so it shows the
vertical lines for THIRDS.
This may be more vivid if you do this yourself, on paper. It should look like this.
In effect we will have eventually turned the initial blank WHOLE into a crossruled grid
— thirds shown vertically, fifths shown horizontally — fifteen small rectangular bricks
altogether in ONE WHOLE; 12 bricks laid in horizontal rows make 4/5, and 10 bricks
stacked in vertical columns make 2/3.
The rest follows fairly easily. The question can be answered by visual inspection, with
shading, pointing and counting, or even placing of actual “bricks” or similar counters.
How many lots of 4/5 (i.e. how many lots of 12 small bricks) are there in 2/3 (i.e. 10
small bricks)?
We can tackle this simply by counting (as we point to parts of the diagram). The answer
is that there are 10 bricks out of 12, or 10 TWELFTHs, which is also the same as 5
SIXTHs.
We can also see this, physically, if we imagine that we take the 2/3 diagram, as shown,
and cut out the small rectangular bricks (each of which has been labelled uniquely with
a letter of the alphabet, so we can see and name which brick is which, without confusing
ourselves), and place them one at a time ON TOP of the 4/5 diagram, brick by brick,
with the final result, as shown,
and then we can see immediately how much of the 4/5 is filled by the 2/3.
This is, naturally, the same answer, TENTWELFTHs, but visually and physically
modelled with a (slightly) different, but related process or mathematical reasoning.
The formal PROOF that invertandmultiply is the correct rule requires:
• some algebra: that is being comfortable with reasoning mathematically using
general letters or PRONUMERALS and standard arithmetic operations, +, –, x,
and ÷, and brackets, ( , and ), instead of using specific numbers and arithmetic;
and
• good understanding of the idea of “fraction cancelling”: here “cancelling” is based
on the ideas of:
— associativity of multiplication: that is, that when we multiply three numbers,
such as F, G and H, it does not matter if we multiply F and G first, and then
multiply by H, or multiply G and H, and then multiply F: that is, algebraically
(FxG)xH = Fx(GxH); and
— the multiplicative identity, 1: this is the special property of 1, that any number
multiplied by ONE is not altered by the multiplying, or 1xN = N, for any
number N — that is, when multiplied by 1, any number N retains its numerical
“identity”; and
— equivalent fractions: such as 1/2 being the same numerical size as 2/4, or
50/100.
Johnson and Rising (1972 pp. 146147) also suggest much more abstract algebraic
approaches to demonstrate that this approach works in ALL cases, not just the caseby
case examples I have modelled here. For any whole numbers, A, B, and C:
if A ÷ B = C, then
(by definition of division, and the fact that dividing is a reverse of multiplying) we know
that
A = C x B
We use this as follows, replacing the whole numbers A, B and C, with fractions or
rational numbers, such as m/n, and r/s, where m, n, r and s are whole numbers, and n
and s are NONzero (because dividing by a zero is not defined, and hence is forbidden).
That is, if the A we start with is a fraction, such as (m/n), and the B we are dividing by
is also a fraction, such as (r/s): and we then divide m/n by r/s, we will have some
resulting number c, which we can work out to be a sensible fraction: that is we start with
—
m ÷ r = c or
n s
Remembering that if A ÷ B = C, then A = C x B, we exchange the A and B parts for the
fractions we are working with, and see that
m = c x r or
n s
This is then further manipulated, using simple arithmetic, to yield the invertand
multiply situation.
We do this by multiplying both sides of the last equation by s/r, and simplifying both
sides of the equal sign, and obtaining a result equal to c.
m x s = ( c x r ) x ( s )
n r s r
The reason we can simplify is that we can see that r/s multiplied by s/r is equal to 1, and
multiplying by 1 does not change the number being multiplied.
This may be more familiar if we talk about “cancelling” the r in the top and the r in the
bottom.
In fact the mathematical argument about cancelling is that we can reorganise the
multiplications in top and bottom, so that we align the r and s parts, and then we have r/r
and s/s and we can see that each is equal to 1, and multiplying by 1 does not alter the
number being multiplied.
The argument is finished by equating the two expressions that have been shown to equal
c.
That is:
In other words, dividing m/n by a fraction r/s is the same as multiplying m/n by s/r.
Voila! We have inverted, and multiplied, in all cases!
Johnson and Rising also offer a general algebraic argument based more directly on the
multiplicative identity for any number a, and the special number 1, namely that,
a x 1 = a.
They rewrite the division of two fractions
(m/n) ÷ (r/s)
as a fraction whose numerator and denominator are both fractions
(m/n) or
(r/s)
They multiply this by 1, and, cunningly, reexpress this 1 as:
(s/r) / (s/r).
Written carefully as numerator vertically (visually) above denominator, this looks like
this:
and they manipulate and simplify, as follows.
When I started writing this explanation, I knew there was an abstract way of proving the
rote invertandmultiply rule. But I lacked the brains to construct it, on my own, and
lacked the longterm memory of it that might have helped me to reconstruct it from
memory.
Fortunately (or not, if you haven't followed this discussion), I happened to have an old
howtoteachmathematics textbook, that provided the information I needed.
The fact that this algebraic proof is so abstract, and thereby hard to understand, and also
so hard to remember, means that we find ourselves forced to reply on the rote rule, or
resorting to clumsy diagrams. I hope some of this is helpful,
John