Steps / Historical AI

Leonhard Euler · 1770 (1828 English ed.) · Internet Archive

Elements of Algebra

All 627 leaves · leaf 536

Transcription

CHAP. XI. ADDITIONS. 499 Now, as we must have $pq' - qp' = \pm 1$, we shall also have, by substituting the preceding values of $p$ and $q$, and effacing what is destroyed, $p'q'' - q'p'' = \mp 1$. Again, substituting in this equation the values of $p'$ and $q'$, there will result $p''q''' - q''p''' = \pm 1$, and so on; so that we shall have, generally, $pq' - qp' = \pm 1$ $p'q'' - q'p'' = \mp 1$ $p''q''' - q''p''' = \pm 1$ $p'''q^{iv} - q'''p^{iv} = \mp 1, \ldots$ So that, if $q'''$, for example, were $= 0$, we should have $-q''p''' = \pm 1$; also, $q'' = 1$, and $p''' = \mp 1$: but if $q^{iv}$ were $= 0$, we should have $-q'''p^{iv} = \mp 1$; therefore $q''' = 1$, and $p^{iv} = \pm 1$; so that, in general, if $q^\varrho = 0$, we shall have $q^{\varrho-1} = 1$; and then $p^\varrho = \pm 1$, if $\varrho$ is even, and $p^\varrho = \mp 1$, if $\varrho$ is odd. Now, as we do not previously know whether the upper, or the under sign is to take place, we must successively suppose $p^\varrho = 1$, and $= -1$: but I say that one of these cases may at all times be reduced to the other; and, for this purpose, it is evidently sufficient to prove, that we can always make the $\varrho$ of the term $q^\varrho$, which must be nothing, either even, or odd, at pleasure. For example, let us suppose that $q^{iv} = 0$, we shall then have $q''' = 1$, and $q'' > 1$, that is, $q'' = 2$, or $> 2$, because the numbers $q, q', q'', \ldots$ naturally form a decreasing series; therefore, since $q'' = \mu'''q''' + q^{iv}$; we shall have $q'' = \mu'''$, so that $\mu''' = 2$ or $> 2$; thus, if we choose, we may diminish $\mu'''$ by unity, without that number being reduced to nothing, and then $q^{iv}$, which was $0$, will become $1$, and $q^v=0$; for putting $\mu''' - 1$, instead of $\mu'''$, we shall have $q'' = (\mu''' - 1)q''' + q^{iv}$; but $q'' = \mu'''$, $q''' = 1$; wherefore, $q^{iv} = 1$; then having $q''' = \mu^{iv}q^{iv} + q^v$, that is, $1 = \mu^{iv} + q^v$, we shall necessarily have $\mu^{iv} = 1$, and $q^v = 0$. Hence we may conclude, in general, that if $q^\varrho = 0$, we shall have $q^{\varrho-1} = 1$, and $p^\varrho = \pm 1$, the doubtful sign being arbitrary. Now, if we substitute the values of $p$ and $q$, given by the preceding formulæ, in $p - aq$, those of $p'$ and $q'$, in $p'-aq'$, and so of others, we shall have $p - aq = \mu (p' - aq') + p'' - aq''$ $p' - aq' = \mu' (p'' - aq'') + p''' - aq'''$ $p'' - aq'' = \mu'' (p''' - aq''') + p^{iv} - aq^{iv}$ $p''' - aq''' = \mu''' (p^{iv} - aq^{iv}) + p^v - aq^v, \ldots$ whence we find KK 2
confidence 0.94 69 expressions source scan at the Internet Archive

This text was produced by machine and has not been corrected by hand. Where it matters, read the scan. Confidence is the model's own, not an external check.