Permissible Acts in Prayer and When Prayer May or Must Be Stopped

1. Actions the Source Does Not Treat as Makruh

The source separates several harmless acts from the Makruh list. Examples include:

• fastening a belt;
• carrying a sword or similar item when its movement is not distracting;
• facing a hanging Mushaf, a hanging sword, the back of a seated person, a candle, or a lamp;
• praying on a rug or mat;
• prostrating on a rug containing pictures when the prostration is not on the picture itself;
• repeating the same Surah in two rak'ahs of Nafl prayer.

The point is not that every movement or object is harmless; rather, these specific cases are not treated as Makruh in the stated rule.

2. Small Necessary Adjustments and Harmful Creatures

Other permitted cases include:

• lightly shaking clothing so it does not cling to the body in Ruku;
• wiping dust or grass from the forehead after prayer;
• wiping it during prayer when it causes harm or serious distraction;
• looking from the corners of the eyes without turning the face;
• killing a snake or scorpion when harm is feared.

The source does not set a fixed maximum of 3 strikes for killing a dangerous creature. It expressly allows multiple blows where needed. A turn of the chest away from the Qiblah remains a separate invalidating issue.

3. Looking at Writing Is Not the Same as Reading from a Mushaf

Looking at written words and understanding them silently does not by itself invalidate prayer in the rule taught here.

This must be kept separate from the earlier technical invalidator concerning reciting from a Mushaf what one has not memorized under Abu Hanifah's position.

So the questions are different:
1. Did the person merely look and understand silently?
2. Or did the person actually use the written text as a source of recitation?

The course now keeps these two issues separate.

4. When Stopping the Prayer Becomes Obligatory

The source makes an important distinction between must stop and may stop.

Prayer must be stopped when an urgently distressed person calls the worshipper for help and the situation genuinely requires rescue.

A parent merely calling — without urgent distress — does not by itself require a person to break a Fard prayer. The commentary gives a separate, more flexible rule for Nafl when a parent does not know the child is praying.

Likewise, if a blind person is very likely to fall into a well or comparable danger, the commentary upgrades intervention from merely permissible to obligatory.

5. When Stopping the Prayer Is Permitted

The source permits stopping the prayer in cases such as:

• preventing theft of property worth the classical threshold of one dirham, even if the property belongs to someone else;
• protecting livestock from a wolf or comparable immediate attack;
• preventing a blind person from falling into a well or a similar danger.

These permissions respond to real danger to life or protected property. They are not a general license to interrupt prayer for ordinary inconvenience.

6. Midwife and Traveler: Delaying Prayer in an Emergency

The same chapter also discusses emergency delay:

• If a midwife fears that a child may die, she gives the necessary care priority. If already praying, the prayer is stopped; if the emergency exists before starting, the prayer may have to be delayed while the child is attended.
• A traveler who genuinely fears robbers or highway attackers is given an emergency allowance to delay the current prayer under the stated rule.

These are exceptional danger cases, not ordinary scheduling preferences.

7. No Invented 'Three-Movement' or 'Three-Scratch' Rule

The previous version contained fixed-number shortcuts about striking or scratching. Those have been removed.

The source does not say:
• “a dangerous creature may be struck only 3 times,” or
• “scratching more than 2 or 3 times in one rak'ah automatically invalidates prayer.”

The real framework is the one already studied:
• necessary small action can be allowed;
• pointless small action can be Makruh;
• Amal Kathir can invalidate prayer.

A fixed number should not be invented where the source does not give one.