Breaking an equation cluster that does not fit on the rest of page across two pages

I wonder why the compiler shift the item that does not fit the rest of the page to the next pages ?! How to tell the compiler to break such item ?!

Here is a screenshot

and here is code regarding this part

\documentclass[12pt,a4paper]{article}
\usepackage[utf8]{inputenc}
\usepackage{amsmath}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsthm}
\usepackage[utf8]{inputenc}
\usepackage{titlesec}
\usepackage{titletoc}
\usepackage{enumitem}
\usepackage{mathtools}
\setlength{\footskip}{15pt}

\begin{document}

XXX\\[17.5cm]
YYY
\begin{align*}
[\textbf{a},\textbf{b}]_{\mathbb{R}^9}
= \phi\bigl([\phi^{-1}(\textbf{a}),\phi^{-1}(\textbf{b})]_{\mathfrak{u}(3)} \bigr) .
\end{align*}

We check the verification of the conditions the bracket $[\,\, , \,\,]_{\mathbb{R}^9}$ must satisfy in the following three steps:

\begin{enumerate}[nolistsep]
\item[(i)] (The $\mathbb{R}$-linearity of $[\,\, ,\,\,]_{\mathbb{R}^9}$) For arbitrary $\textbf{a},\textbf{b},\textbf{c} \in \mathbb{R}^9$ and $\alpha_1,\alpha_2 \in \mathbb{R}$, it follows by the $\mathbb{R}$-linearity of $[\,\, \,\,]_{\mathfrak{u}(3)},\phi$ and  $\phi^{-1}$ that
\begin{equation*}
\begin{split}
\bigl[\alpha_1 \textbf{a} + \alpha_2 \textbf{b} , \textbf{c}    \bigr]_{\mathbb{R}^9}
&= \phi\bigl( [\phi^{-1}(\alpha_1\textbf{a}+\alpha_2 \textbf{b}), \phi^{-1}(\textbf{c})]_{\mathfrak{u}(3)}  \bigr)
\\&= \phi\bigl( [\alpha_1 \phi^{-1}(\textbf{a}) + \alpha_2 \phi^{-1} (\textbf{b}), \phi^{-1}(\textbf{c}) ]_{\mathfrak{u}(3)}    \bigr)
\\&= \phi\bigl( \alpha_1 [\phi^{-1}(\textbf{a}), \phi^{-1}   (\textbf{c})]_{\mathfrak{u}(3)} + \alpha_2 [\phi^{-1}(\textbf{b}), \phi^{-1}  (\textbf{c})]_{\mathfrak{u}(3)}       \bigr)
\end{split}
\end{equation*}

\end{enumerate}

\end{document}


latex split the block of equations normally except that when I use enumerate environment latex shifts the item entirely to the next page.. This is the only exception and I unfortunately do not know why !

• Have you tried providing the instruction \allowdisplaybreaks in the preamble (after loading amsmath)? – Mico Dec 3 '16 at 23:20
• as I said in the question you just deleted the example code is only useful if it demonstrates the problem. The example here doesn't show a bad page break. If it was a break Mico's answer should work but no one can test as you have provided no example, but looking at your image it is not at all clear there is space for any more text on the first page. – David Carlisle Dec 3 '16 at 23:31
• No you have not provided any useful code. You are asking about a bad page break so your code should demonstrate a bad page break, how can anyone test answers given a test case that fits on one page? – David Carlisle Dec 3 '16 at 23:33
• also you should say what output you expect. I can not see anything wrong with the images that you have posted (but as they are not made by the posted code they are not very helpful) – David Carlisle Dec 3 '16 at 23:36
• That still wont show an actual page break. Can you either \vspace or some nonsense text so that when we compile it we can see exactly what you have in the screenshot? – Torbjørn T. Dec 3 '16 at 23:44

Using align*and allowdisplaybreaks works quite normally (I simplified your code for the enumerate environment). \allowdisplaybreaks has to be used before the equation environment we want to break (I put it in the preamble, so all multilined equations will be broken if necessary):

\documentclass[12pt,a4paper]{article}
\usepackage[utf8]{inputenc}
\usepackage[showframe]{geometry}%
\usepackage{lipsum}% \usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsthm}
\usepackage[utf8]{inputenc}
\usepackage{titlesec}
\usepackage{titletoc}
\usepackage{enumitem}
\usepackage{mathtools}
\setlength{\footskip}{15pt}
\allowdisplaybreaks

\begin{document}

\lipsum[1-4]

\begin{enumerate}[label=(\roman*), noitemsep, topsep=2pt]
\item (The $\mathbb{R}$-linearity of $[\,\, ,\,\,]_{\mathbb{R}⁹}$) For arbitrary $\textbf{a},\textbf{b},\textbf{c} ∈ \mathbb{R}⁹$ and $\alpha₁,\alpha₂ ∈ \mathbb{R}$, it follows by the $\mathbb{R}$-linearity of $[\,\, \,\,]_{\mathfrak{u}(3)},\phi$ and $\phi⁻¹$ that
\begin{align*}
\bigl[\alpha₁ \textbf{a} + \alpha₂ \textbf{b} , \textbf{c} \bigr]_{\mathbb{R}⁹}%
& = ϕ\bigl( [\phi⁻¹(\alpha₁\textbf{a}+\alpha₂ \textbf{b}), \phi⁻¹(\textbf{c})]_{\mathfrak{u}(3)} \bigr) \\
& = ϕ\bigl( [\alpha₁ \phi⁻¹(\textbf{a}) + \alpha₂ \phi⁻¹ (\textbf{b}), \phi⁻¹(\textbf{c}) ]_{\mathfrak{u}(3)} \bigr) \\ %
& = ϕ\bigl( \alpha₁ [\phi⁻¹(\textbf{a}), \phi⁻¹ (\textbf{c})]_{\mathfrak{u}(3)} + \alpha₂ [\phi⁻¹(\textbf{b}), \phi⁻¹ (\textbf{c})]_{\mathfrak{u}(3)} \bigr)
\end{align*}

\item \lipsum[5] %
\end{enumerate}

\end{document}


• what does \lipsum[] stand for ? – Hussein Eid Dec 4 '16 at 0:02
• I got this error " Missing number, treated as zero" @Mico – Hussein Eid Dec 4 '16 at 0:06
• @HusseinEid - Why are you addressing this comment at me? – Mico Dec 4 '16 at 0:07
• @Hussein Eid: It's a command to produce automatically fake text is some pseudo Latin language (the optional argument is the numbering of the paragraphs we want to obtain). Of course you have to use it: it's just here to produce a page break in the middle of the group of equations. – Bernard Dec 4 '16 at 0:08
• @Mico: Done, mylord! – Bernard Dec 4 '16 at 0:11